<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Prabu Sangar on Medium]]></title>
        <description><![CDATA[Stories by Prabu Sangar on Medium]]></description>
        <link>https://medium.com/@prabu.sangar?source=rss-ed44d3335f1d------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*SObhhtGoKR39maJzCTj46w.png</url>
            <title>Stories by Prabu Sangar on Medium</title>
            <link>https://medium.com/@prabu.sangar?source=rss-ed44d3335f1d------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 21:36:36 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@prabu.sangar/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Lying to Belong]]></title>
            <link>https://medium.com/@prabu.sangar/lying-to-belong-8baa8bb6e649?source=rss-ed44d3335f1d------2</link>
            <guid isPermaLink="false">https://medium.com/p/8baa8bb6e649</guid>
            <category><![CDATA[social-media]]></category>
            <category><![CDATA[lying]]></category>
            <category><![CDATA[psychology]]></category>
            <category><![CDATA[sociology]]></category>
            <category><![CDATA[self-reflection]]></category>
            <dc:creator><![CDATA[Prabu Sangar]]></dc:creator>
            <pubDate>Sat, 08 Aug 2026 01:30:44 GMT</pubDate>
            <atom:updated>2026-08-08T01:31:33.451Z</atom:updated>
            <content:encoded><![CDATA[<h4>Is pathological lying becoming normal, or are we simply trying harder to become people others will accept?</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*ZKBxLPRUS23EqeiT" /><figcaption>Photo by <a href="https://unsplash.com/@reskp?utm_source=medium&amp;utm_medium=referral">Jametlene Reskp</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h3>Self-Reflection</h3><p>I have been thinking about lying lately, although perhaps not in the way we normally think about it. I am not talking about fraud, affairs, criminal deception or the dramatic lies that destroy families and careers. I am thinking about the smaller lies we tell when we want to belong, when the truth feels slightly less impressive than what the room expects from us, or when we quietly edit ourselves because we worry that the unedited version may not be enough.</p><p>I struggle with this subject because I cannot honestly place myself completely outside it. There have been moments when I have made something sound better than it was, withheld an uncomfortable detail, exaggerated something slightly or presented the version of myself that I thought would be better received. Sometimes I knew exactly what I was doing. At other times, I realised only afterwards that I had adjusted the story without really questioning why.</p><p>I also see versions of this around me. We polish professional histories, describe difficult relationships differently depending on who is listening, make achievements sound slightly larger and occasionally allow failures to disappear from the stories we tell about ourselves. Social media makes this even easier because we can construct the version of our lives that other people are allowed to see.</p><p>Reality check and a reminder for all of you. None of this means that I, or everyone doing these things, is a pathological liar. That distinction is important because psychological language has become dangerously casual. Someone lies to us and we call them pathological. Someone behaves selfishly and they become a narcissist. Someone manipulates a situation and suddenly they are a psychopath or sociopath.</p><blockquote>Psychology is considerably more complicated.</blockquote><p>The question I am interested in is not simply why people lie. Human beings have always lied. I am more interested in why we sometimes feel the truth needs improving before another person can accept it, and whether modern society is encouraging a form of identity management in which small distortions have become so ordinary that we barely recognise them as dishonesty.</p><h3>Most People Are Probably Not Lying</h3><p>There is a popular belief that everyone lies constantly. The research presents a more interesting picture.</p><p>One major study of 1,000 American adults found an average of 1.65 lies per person per day, which sounds as though dishonesty is routine until the distribution is examined properly. Almost 60 per cent of participants reported telling no lies during the previous 24 hours, while just 5.3 per cent of participants accounted for half of all reported lies. The average therefore concealed an important reality. Most people were relatively honest, while a small number of people were responsible for a disproportionately large amount of deception.</p><p>A later study followed 632 people for 91 days and recorded 116,366 lies. Approximately three quarters of participants were consistently low-frequency liars, while 5.7 per cent were classified as prolific liars who averaged six or more lies per day. That small group was responsible for 28 per cent of all lies recorded during the study. Most lies across the entire sample were also relatively minor white lies rather than elaborate acts of deception.</p><p>This matters because it challenges my original question. Pathological lying may not be becoming the new norm at all. We may instead be living in a society where a relatively small number of prolific liars coexist with a much larger population that participates in occasional impression management.</p><blockquote>Those are not the same psychological phenomenon.</blockquote><h3>What Pathological Lying Actually Means</h3><p>Pathological lying has been discussed in medical and psychiatric literature for more than a century, sometimes under the term pseudologia fantastica. Despite its long history, it remains poorly understood and is not currently recognised as an independent psychiatric diagnosis in standard diagnostic systems.</p><p>Researchers Drew Curtis and Christian Hart proposed that pathological lying involves a persistent, pervasive and often compulsive pattern of excessive lying that produces significant distress, impaired functioning or danger and continues for longer than six months. In their study, 13 per cent of participants reported that either they themselves or people who knew them regarded them as pathological liars. Those individuals reported more distress, poorer functioning and greater danger associated with their lying than other participants. Their lies also appeared to become compulsive, with additional fabrications developing from an original lie even when there was no obvious reason to continue.</p><p>The 13 per cent figure needs to be treated carefully. It does <strong>not</strong> mean that 13 per cent of the population has pathological lying disorder. The study was not designed to establish a national or global prevalence rate, and pathological lying is not currently a formally diagnosed disorder. It is better understood as evidence that a meaningful minority of people report a pattern of lying sufficiently severe to cause distress and impairment.</p><p>A 2026 systematic review also shows how limited the evidence remains. Researchers identified only 34 published articles containing 55 documented cases of pseudologia fantastica. Among those cases, common reported motives included self-aggrandisement, attention seeking and assuming the sick role. Case reports are useful for understanding clinical patterns, but 55 published cases cannot tell us how common the behaviour is in the wider population.</p><p>In my opinion, this uncertainty is important because we have become much more confident in using psychological labels than psychology itself is in defining them.</p><h3>We Want To Belong</h3><p>One of the most useful ways of understanding everyday dishonesty comes from Baumeister and Leary’s <strong>Need to Belong hypothesis</strong>. Their argument was that human beings possess a powerful motivation to form and maintain stable interpersonal relationships, and that threats to belonging have significant emotional and behavioural consequences. Belonging is not merely pleasant. It appears to be a fundamental part of psychological functioning.</p><p>This theory changed the way I thought about some ordinary lies.</p><p>If belonging matters deeply to us, then lying can sometimes function as a social adaptation. We may pretend to know something because everyone else seems knowledgeable. We may exaggerate an achievement because we want respect. We may conceal financial difficulty because our social circle appears successful. We may claim to enjoy something because disagreement feels socially expensive.</p><p>The lie is not necessarily motivated by a desire to exploit anyone. Sometimes it is motivated by the fear that the truth will make us less acceptable.</p><p>I recognise this in myself. There have been occasions when I have wanted to sound more certain than I actually was because uncertainty felt uncomfortable in that particular room. There have probably also been moments when I presented a cleaner version of a difficult situation because explaining the full mess felt embarrassing.</p><blockquote>The psychologically uncomfortable question is whether I was lying to other people or protecting myself from what I imagined they might think of me. Perhaps both were happening.</blockquote><h3>Person We Think We Should Be</h3><p>E. Tory Higgins’ <strong>Self-Discrepancy Theory</strong> provides another explanation. Higgins distinguished between the actual self, the person we believe we currently are, the ideal self, the person we hope or wish to become, and the ought self, the person we believe we are expected to be. Distress can emerge when these versions move too far apart.</p><p>Social media may have created an almost perfect environment for amplifying these discrepancies.</p><p>The actual self may be exhausted, uncertain about work, worried about money and experiencing an ordinary relationship with ordinary problems. The ideal self is travelling, exercising, progressing professionally, emotionally fulfilled and apparently ageing very well. The ought self adds another layer because society tells us where we should be by a certain age, what success should look like and what a good relationship, career or lifestyle should resemble.</p><p>The distance between these identities can become uncomfortable. A small lie provides a temporary shortcut.</p><p>Instead of becoming the person I wish I were, I can describe myself as though I am already there. The psychological gap closes for a moment, particularly when other people respond positively.</p><blockquote>That is where the behaviour can become reinforcing.</blockquote><h3>Impression Management Model</h3><p>Leary and Kowalski’s <strong>Impression Management Model</strong> helps explain why editing ourselves is not automatically pathological. Their model proposes that people become motivated to control how others perceive them when the impression matters to an important goal, when the desired outcome is valuable and when there is a discrepancy between how they currently believe they are seen and how they want to be seen. They then construct an image based partly on their identity, social role, audience and desired outcome. Most of us do this every day.</p><p>I do not speak exactly the same way in a professional meeting as I do with close friends. I would not describe myself in a job interview using every insecurity, failure and questionable decision I have made throughout my career. That is not lying. It is social competence. The problem begins when impression management requires fabrication.</p><p>A recent 2026 study examining social-media self-presentation built directly on Higgins’ self-discrepancy theory, Goffman’s ideas about self-presentation and Leary’s impression-management model. Researchers identified dimensions including reality distortion and exaggeration, idealised self-presentation and strategic impression management. The fact that researchers now need instruments capable of measuring the distance between online presentation and the real self tells us something about where digital identity has taken us.</p><p>Another study examining Instagram found that a stronger need for approval was associated with more deceptive self-presentation. False self-presentation was also associated with greater perceived popularity, but at the same time with poorer psychological outcomes including depression.</p><p>That combination interests me because it captures the contradiction perfectly. The edited self may perform better socially while the person behind it feels worse.</p><h3>Social Comparison Moves Target</h3><p>Festinger’s <strong>Social Comparison Theory</strong> argued that when objective standards are unavailable, people often evaluate themselves by comparing themselves with others. The theory was developed long before Instagram or LinkedIn, but it is difficult to imagine a more powerful demonstration of the concept than modern social media.</p><p>We compare careers with people announcing promotions, bodies with people posting selected photographs, relationships with anniversary messages and lifestyles with holidays. The comparison itself is already unfair because we are comparing the complete version of our lives with selected pieces of someone else’s. Then something more interesting happens. We begin editing ourselves in response.</p><p>If everyone appears successful, I become more tempted to highlight success. If everyone appears happy, unhappiness becomes something I may want to hide. If professional progress appears effortless, admitting that I am struggling can feel like evidence that I am falling behind.</p><p>Everyone becomes both performer and audience. The more people curate themselves, the more pressure everyone else experiences to curate themselves too. We may therefore be participating in a system where nobody deliberately created the standard, yet everyone contributes to maintaining it. That is why I hesitate to describe the issue simply as dishonesty.</p><blockquote>Sometimes the lie is part of a comparison system in which everyone thinks everyone else is doing better than they actually are.</blockquote><h3>Cognitive Dissonance</h3><p>Lying can also create tension inside the person telling the lie. <strong>Cognitive Dissonance Theory</strong>, associated with Leon Festinger, proposes that people experience psychological discomfort when their actions, beliefs and self-concepts conflict. Someone who considers themselves an honest person but repeatedly behaves dishonestly has to manage that contradiction somehow. One option is to stop lying. Another is to change the explanation.</p><p>I may tell myself that everybody exaggerates, that the lie was harmless, that the other person deserved it, that I had no choice or that the statement was technically true if interpreted in a particular way. Each explanation reduces the discomfort between the person I believe I am and the behaviour I have actually displayed.</p><p>This does not mean every liar experiences obvious guilt. The relationship between deception, personality and emotional response is complicated. However, recent research involving four studies and 783 participants found that lying was associated with lower self-esteem and more negative emotion, including when people recalled previous lies and when researchers followed changes within individuals over time.</p><p>I find this relevant to my own smaller experiences with dishonesty. There is often a moment after exaggerating something when I wonder why I felt the need to do it in the first place. The lie may have improved the story but somehow made me less comfortable with the person telling it.</p><p>Perhaps this is because approval gained through an inaccurate version of ourselves never fully belongs to us.</p><h3>Rewarding Lies</h3><p>Behavioural psychology provides a simpler explanation for why some forms of lying can become repetitive. A behaviour that receives a reward is more likely to be repeated.</p><p>Imagine someone exaggerates a story at dinner and everyone laughs. The person gains attention. Someone invents professional success and receives admiration. Someone makes themselves the victim in a conflict and receives sympathy. Someone presents a highly edited life online and receives hundreds of positive reactions. The immediate consequences are rewarding. The brain does not necessarily conduct an ethical review before learning from reinforcement.</p><p>If deception repeatedly produces social approval, escape from embarrassment or avoidance of punishment, the behaviour can become easier to repeat. The first lie may require deliberate thought, while later lies begin to emerge more automatically because the behaviour has repeatedly worked.</p><p>This still does not make the behaviour pathological. Many behaviours become habitual without becoming psychiatric disorders. However, it provides one way of understanding how occasional deception can gradually become part of someone’s normal response to social discomfort.</p><h3>Believing Liars</h3><p>Timothy Levine’s <strong>Truth-Default Theory</strong> adds another important piece. The theory proposes that people generally assume others are telling the truth unless something triggers sufficient suspicion to question them. This tendency is useful because ordinary human communication would become almost impossible if we treated every sentence as potentially fraudulent.</p><p>Most of the time, believing people works because most communication is truthful. Unfortunately, the same mechanism creates an opportunity for people who lie frequently.</p><p>A convincing liar does not necessarily need extraordinary psychological powers. They are often operating in a social environment where everyone else begins from the assumption that what they are hearing is probably true.</p><p>This may also explain why repeated deception can be so damaging to relationships. Once the truth-default collapses, the relationship can move to the opposite extreme. A partner, friend or colleague who has discovered repeated lies may begin questioning statements that are completely true.</p><blockquote>Trust is therefore not simply lost. The basic operating system of the relationship changes.</blockquote><h3>Psychopathy or Sociopathy?</h3><p>This is where I think popular culture has created unnecessary confusion. Lying does not make someone a psychopath.</p><p>Psychopathy is generally understood as a broader cluster of interpersonal, emotional and behavioural traits that may include manipulation, superficial charm, limited empathy, reduced remorse and pathological lying. It overlaps with antisocial personality characteristics but is not simply another word for someone who lies frequently.</p><p>Sociopathy is even less precise as a clinical term and is often used informally when people are discussing antisocial personality disorder. The motivation behind the lie therefore matters.</p><p>Someone with strong psychopathic or antisocial traits may deceive another person deliberately because the lie provides money, control, status or some strategic advantage. Someone engaging in pathological lying may continue creating stories when the practical benefit is unclear and the lying itself is causing serious personal damage.</p><p>Then there is the person who lies because they are ashamed, insecure or desperate to belong. The behaviour may look similar from the outside while the psychology underneath it is completely different. This is one reason we should be careful about diagnosing people from their worst behaviour.</p><p>Explanation does not remove accountability. If someone repeatedly lies, manipulates or harms another person, psychological explanations do not require the victim to tolerate that behaviour. Understanding why someone behaves a particular way and deciding what behaviour we are willing to accept are separate questions.</p><h3>Tragedy of Misidentity</h3><p>I think one of the saddest possibilities surrounding chronic lying is that a person can eventually receive exactly the admiration they wanted and still be unable to experience it fully.</p><p>Imagine inventing an achievement and watching people respond with respect. The recognition may initially feel good, but the person also knows something nobody else in the room knows. The admiration belongs to an identity that does not exist.</p><p>The more successful the invented identity becomes, the more expensive honesty becomes. One lie then requires supporting details. Questions create more explanations. New relationships require the story to be repeated. Contradictions need to be repaired and different audiences may know different versions of the same person. Eventually, significant psychological effort may be spent maintaining a reality that never existed. At that stage, the lie is no longer simply protecting the person. The person is protecting the lie.</p><h3>It’s Everywhere</h3><p>This is the part that makes the subject uncomfortable for me because I can see smaller versions of these mechanisms in completely ordinary behaviour.</p><p>Careers are presented as though every decision was strategic, even though most careers contain accidents, compromises, mistakes and periods where nobody knew exactly what they were doing. Relationships are presented differently depending on the audience. Achievements become slightly more impressive after several retellings, while embarrassing details slowly disappear.</p><p>I have caught myself doing smaller versions of this as well, which is why I find it difficult to write about lying purely as a moral defect. Sometimes dishonesty is deliberate manipulation and should be treated accordingly, but sometimes I think the underlying emotion is fear. We want the person in front of us to approve of us, and for a few seconds changing the story feels safer than allowing them to judge the truth.</p><blockquote>That does not make the dishonesty acceptable, but it does make the psychology worth examining.</blockquote><h3>Social Media and Pathological Liars</h3><p>I do not think there is sufficient evidence to conclude that pathological lying itself has become a new social norm. What has changed is the infrastructure surrounding self-presentation.</p><p>We now have permanent professional profiles, dating profiles, photographs, follower counts, public career announcements and personal brands. A carefully constructed identity can be presented continuously to people who may never encounter enough of our offline life to challenge it.</p><p>The social reward is also visible. Popularity has numbers attached to it.</p><p>LinkedIn gives us a particularly interesting version because the editing often remains close enough to the truth to avoid being called lying. We rarely see careers described as uncertain or unsuccessful. A redundancy becomes an opportunity to pursue new directions. A failed project disappears. Routine responsibilities can become leadership. A short interaction with an influential person can gradually become a professional relationship. Much of this is simply communication and marketing. Yet there is a point where presentation becomes distortion, and I am not convinced we always know exactly where that point sits. Perhaps social media has not produced an epidemic of pathological lying.</p><blockquote>Perhaps it has produced an epidemic of <strong>identity insecurity</strong>.</blockquote><h3>Ordinary Not Enough</h3><p>The more I think about the psychological theories together, the more they appear to describe different parts of the same process.</p><p>The need to belong explains why rejection matters. Self-discrepancy theory explains the uncomfortable distance between the person we are and the person we think we should be. Social comparison helps explain where that ideal comes from. Impression management explains why we modify our presentation when another person’s opinion matters.</p><p>Reinforcement helps explain why a successful lie may be repeated, while cognitive dissonance helps explain the discomfort and rationalisation that can follow. Truth-default theory then explains why the people around us often believe the story in the first place. None of these theories independently explains pathological lying. Together, however, they help explain why dishonesty can become psychologically attractive. I find that much more useful than simply dividing the world into honest people and liars.</p><h3>Connections</h3><p>The greatest cost of repeated dishonesty may not be getting caught. It may be losing the possibility of being genuinely known.</p><p>Intimacy requires allowing someone else to encounter something real about us. If I constantly provide the version of myself that I believe another person will admire, then even if they eventually admire me, another problem remains. I cannot know whether they would have accepted the person behind the presentation. Lying can therefore become strangely self-defeating. We lie because we want connection, but the lie can make genuine connection impossible.</p><p>Trust also rarely collapses only because of one spectacular deception. Sometimes it disappears through a succession of small inconsistencies until the other person begins questioning everything, including statements that are actually true. At that point, honesty becomes much harder to recover because the relationship no longer has a reliable way to distinguish truth from fiction.</p><h3>Back to My Reflections</h3><p>I started with a question about whether pathological lying was becoming the new norm, but I no longer think that is the right question.</p><p>The available evidence does not demonstrate that society is experiencing an epidemic of pathological liars. Most people appear to lie relatively infrequently, and the research consistently shows that a small proportion of people account for a disproportionately large amount of deception.</p><blockquote>What may have changed is our relationship with presentation.</blockquote><p>We have become experienced editors of our own lives. We select, remove, polish and sometimes exaggerate because being ordinary can feel strangely risky in a world where everyone else appears exceptional. The very concept of social media.</p><p>I have to include myself in that observation because there have been moments when I have wanted to appear more certain, successful or comfortable than I genuinely felt. I suspect many people reading this will recognise something similar in themselves. That does not make us pathological liars.</p><blockquote>It may simply make us human beings responding to belonging, comparison, insecurity and the desire to be valued.</blockquote><p>However, understanding why we do it should not become an excuse for continuing to do it. Perhaps the healthier question is not whether we have ever lied to fit in, because most of us probably have in some form. The more difficult question is whether the people closest to us know the person underneath the presentation, and whether we still believe that person deserves to be accepted.</p><p>Pathological lying may represent an extreme psychological problem, but the fear beneath some ordinary lies may be far more familiar. It is the fear that if people knew the complete truth about who we are, the truth might not be enough.</p><h3>References</h3><p>Baumeister, R. F., and Leary, M. R. (1995). <em>The need to belong, Desire for interpersonal attachments as a fundamental human motivation</em>. Psychological Bulletin, 117(3), 497–529.</p><p>Curtis, D. A., and Hart, C. L. (2020). <em>Pathological lying, Theoretical and empirical support for a diagnostic entity</em>. Psychiatry Research and Clinical Practice, 2(2), 62–69.</p><p>Curtis, D. A., and Hart, C. L. (2022). <em>Pathological lying, Psychotherapists’ experiences and ability to diagnose</em>. American Journal of Psychotherapy, 75(2), 61–66.</p><p>Dike, C. C., Baranoski, M., and Griffith, E. E. H. (2005). <em>Pathological lying revisited</em>. Journal of the American Academy of Psychiatry and the Law, 33(3), 342–349.</p><p>Festinger, L. (1954). <em>A theory of social comparison processes</em>. Human Relations, 7, 117–140.</p><p>Higgins, E. T. (1987). <em>Self-discrepancy, A theory relating self and affect</em>. Psychological Review, 94(3), 319–340.</p><p>Kusci, İ., and Bakır, V. (2026). <em>Developing the Social Media Self-Presentation Discrepancy Scale, Psychological and social implications of self-presentation and impression management strategies</em>. The Journal of Social Psychology.</p><p>Leary, M. R., and Kowalski, R. M. (1990). <em>Impression management, A literature review and two-component model</em>. Psychological Bulletin, 107(1), 34–47.</p><p>Levine, T. R. (2014). <em>Truth-Default Theory, A theory of human deception and deception detection</em>. Journal of Language and Social Psychology, 33(4), 378–392.</p><p>Levine, T. R. (2022). <em>Truth-default theory and the psychology of lying and deception detection</em>. Current Opinion in Psychology, 47, 101380.</p><p>Mun, I. B., and Kim, H. (2021). <em>Influence of false self-presentation on mental health and deleting behavior on Instagram, The mediating role of perceived popularity</em>. Frontiers in Psychology, 12, 660484.</p><p>Preuter, S., and colleagues. (2024). <em>The costs of lying, Consequences of telling lies on the liar’s self-esteem and affect</em>. British Journal of Social Psychology.</p><p>Serota, K. B., Levine, T. R., and colleagues. (2021). <em>Unpacking variation in lie prevalence, Prolific liars, bad lie days, or both?</em> Communication Monographs.</p><p><em>Pseudologia fantastica, A systematic review of sociodemographic features and psychiatric comorbidities</em>. (2026). Systematic review of 34 publications and 55 reported cases.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8baa8bb6e649" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Golden Ratio, A Flawed Formula]]></title>
            <link>https://medium.com/@prabu.sangar/golden-ratio-a-flawed-formula-1720ca12202c?source=rss-ed44d3335f1d------2</link>
            <guid isPermaLink="false">https://medium.com/p/1720ca12202c</guid>
            <category><![CDATA[cosmetic-surgery]]></category>
            <category><![CDATA[beauty-standards]]></category>
            <category><![CDATA[psychology]]></category>
            <category><![CDATA[body-image]]></category>
            <category><![CDATA[asian-culture]]></category>
            <dc:creator><![CDATA[Prabu Sangar]]></dc:creator>
            <pubDate>Sun, 26 Jul 2026 21:26:01 GMT</pubDate>
            <atom:updated>2026-07-26T21:26:01.467Z</atom:updated>
            <content:encoded><![CDATA[<h4>What Asia’s beauty industry reveals about psychology, race and the value placed on women’s faces</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*iKj9su4MBCe3kdsr" /><figcaption>Photo by <a href="https://unsplash.com/@dizzyd718?utm_source=medium&amp;utm_medium=referral">Drew Dizzy Graham</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>I have travelled across Asia enough to notice that beauty does not speak the same language everywhere. In Seoul, it can feel highly intentional, technically refined and closely connected to skincare, entertainment and aesthetic medicine. In Tokyo, the preferred image often appears more restrained, youthful and carefully maintained. In Bangkok, beauty is visible, accessible and closely connected with medical tourism, while in India and the Philippines it is difficult to separate beauty from skin colour, celebrity culture, pageants and social mobility. Singapore feels different again because the money is available, yet whatever intervention takes place is often expected to remain discreet.</p><p>The details may differ, but women across these societies receive a remarkably similar message. They may be considered beautiful, but there is always an opportunity to become more beautiful. Their skin can become clearer, their eyes can appear larger, their noses can become more defined, their jaws can become narrower and their faces can remain younger for longer. The improvement is never complete because every correction creates a new point of comparison.</p><p>The beauty industry did not invent human insecurity, and it would be unfair to suggest that every woman who uses cosmetics or undergoes a procedure must dislike herself. What the industry has done exceptionally well is identify insecurity, connect it with social value and transform personal discomfort into a market that offers an endless series of solutions.</p><h3>A Controlversial Claim</h3><p>Here is my deliberately controversial claim. If female beauty were ranked according to how closely each country’s dominant commercial ideal follows facial symmetry, proportion, balance and the measurements commonly associated with the golden ratio, South Korea would come first, followed by Japan and China. Thailand would take the next position because of the visibility of its beauty and aesthetic medicine industries, followed by Australia as a regional comparison, Vietnam, the Philippines, India, Singapore and Malaysia.</p><p>Australia is included as an Asia Pacific comparison rather than as an Asian country. Its inclusion is useful because it allows us to consider whether a wealthy, developed and predominantly Western society produces the same beauty behaviours and pressures seen across East and Southeast Asia.</p><p>Before this ranking becomes the reason someone stops reading, I believe it is important to explain what it does and does not represent. I am not claiming that millions of women across these countries have been scientifically measured and that Korean women are biologically more attractive than Japanese, Chinese, Indian, Thai, Filipino or Australian women. No credible evidence could support that conclusion, and the ethnic and individual diversity within every country would make such an exercise meaningless.</p><p>What I am ranking is the influence of each country’s commercial beauty system and the degree to which its most visible female ideal resembles the proportions celebrated by golden-ratio facial analysis. I am considering the normalisation of aesthetic treatment, the consistency of the female image promoted through media and advertising, the international reach of that image and the extent to which beauty is associated with youthfulness, status, discipline and self-improvement.</p><p>It remains a subjective cultural argument rather than a scientific measurement. That may also be why it matters. Most national beauty rankings claim to measure women when they are really measuring the influence of media, wealth, race, entertainment, aesthetic medicine and economic power.</p><h3>Why South Korea Comes First</h3><p>South Korea comes first not because Korean women possess a superior biological formula, but because the country has created the region’s most recognisable and exportable female beauty ideal. Korean entertainment, skincare, fashion, social media and aesthetic medicine reinforce one another, creating a complete ecosystem in which beauty is not merely admired but continuously managed.</p><p>The dominant commercial image often emphasises clear and even skin, facial balance, a youthful appearance, larger-looking eyes, a defined nose and a narrower lower face. These characteristics are not present in every Korean woman and are not universally preferred within Korea itself, but they appear often enough in entertainment and advertising to have become internationally recognisable as part of the Korean beauty aesthetic.</p><p>The scale of South Korea’s international influence became even clearer through the country’s 2025 medical tourism data. Official government figures released in 2026 showed that more than 2 million foreign patients received treatment in South Korea during 2025, the first time the annual number exceeded that level. Dermatology accounted for approximately 1.3 million foreign patients and plastic surgery accounted for more than 230,000.</p><p>These figures should not be interpreted as proof that every foreign patient travelled to Korea because of insecurity, nor does every dermatological treatment exist solely for cosmetic purposes. They do, however, demonstrate how effectively South Korea has connected medical care, beauty, tourism, entertainment and national influence. Korea no longer simply exports products or celebrities. It exports an idea of what a desirable face can look like and, more importantly, offers the treatments through which a person may attempt to achieve it.</p><p>What makes the Korean system psychologically powerful is the language surrounding it. Cosmetic intervention can be presented less as a rejection of the self and more as a disciplined investment in the self. Surgery becomes refinement, injectables become maintenance and dermatology becomes part of wellness. Once the language changes, the emotional threshold for treatment also changes because a procedure no longer feels like an extreme act. It begins to feel like the responsible management of one’s appearance.</p><h3>The “Universal” Formula</h3><p>The golden ratio is approximately 1.618 and is a legitimate mathematical relationship found in geometry and certain natural patterns. Over time, it also became associated with art, architecture, human proportions and facial beauty. Contemporary facial-mapping systems compare measurements involving the eyes, nose, lips, chin and overall shape of the face before producing a percentage that supposedly indicates how close a person is to mathematical perfection.</p><p>The concept is appealing because it removes uncertainty. Instead of admitting that beauty is subjective, cultural and influenced by the observer, the golden ratio promises that attractiveness can be measured with the authority of mathematics. A person is no longer simply told that someone does not find her beautiful. She is told that her eyes, nose or jaw have failed an objective calculation.</p><p>The scientific foundation for that claim is weak. A major review of the available evidence found no convincing proof that the golden ratio determines ideal human proportions or facial beauty. The researchers also found no evidence supporting its use in planning facial surgery or evaluating surgical outcomes.</p><p>This creates an important contradiction. We continue to use a formula that does not reliably describe natural human beauty, while allowing it to influence how natural faces are judged and sometimes altered. Mathematics is not revealing a universal standard in this situation. A cultural preference is borrowing the language of mathematics so that it appears more credible.</p><h3>Prejudice Acquires a Scientific Appearance</h3><p>The golden ratio becomes particularly concerning when facial proportions derived from narrow populations are applied universally. A measurement may appear neutral, but the construction of the ideal face is never neutral. Someone decided which faces would be measured, which features would be considered balanced and which population would represent the norm.</p><p>This is where the formula can become racism wearing a lab coat. The mathematics itself is not racist, but its application can become racialised when ethnic characteristics are presented as departures from an ideal rather than as ordinary forms of human variation. Wider noses, darker skin, single eyelids, rounder faces and different facial profiles can gradually be interpreted as problems when the supposedly perfect face repeatedly favours a narrower range of characteristics.</p><p>Not every person who undergoes eyelid surgery is trying to look Western, and it would be intellectually lazy to describe every rhinoplasty or skin treatment as racial self-rejection. Beauty standards in China, Japan, South Korea, India and Southeast Asia have their own histories, class structures and cultural preferences. Contemporary ideals are often hybrids created from local traditions, global fashion, celebrity culture, digital filters and Western influence.</p><p>However, personal motivations do not erase social patterns. When darker skin is repeatedly lightened, ethnic features are repeatedly narrowed and a limited facial structure is repeatedly described as more professional, refined or feminine, individual choices begin to reveal the hierarchy surrounding them. A woman may freely choose a treatment while still responding to pressures that she did not personally create.</p><p>That distinction allows us to respect women’s agency without pretending that their choices take place in a cultural vacuum.</p><h3>Latest Data Reflection</h3><p>The latest internationally comparable survey from the International Society of Aesthetic Plastic Surgery was released in 2025, although it measured procedures performed during 2024. A complete global survey for procedures performed during 2025 has not yet been published, so it would be misleading to relabel older worldwide estimates as 2025 data.</p><p>The latest report estimated that approximately 17.4 million surgical and 20.5 million non-surgical aesthetic procedures were performed globally. Japan recorded approximately 1.63 million procedures, of which more than 1.25 million were non-surgical. India recorded approximately 1.29 million procedures, while Vietnam recorded about 262,000, Thailand about 197,000 and Australia about 177,000. Malaysia, the Philippines and Singapore recorded considerably smaller raw totals.</p><p>China and South Korea did not have separately published national procedure totals because the required survey response threshold was not met, although the report estimated that both countries had substantial numbers of plastic surgeons. Their absence from the country tables should not be interpreted as evidence that cosmetic intervention is uncommon. It simply means that the available data was not strong enough to produce a statistically valid national estimate.</p><p>These figures should not be treated as an obsession ranking. They count procedures rather than individual patients, meaning that one person receiving several treatments may be counted several times. The results are also affected by population size, medical tourism, differences in reporting and whether treatments are provided by plastic surgeons, dermatologists or practitioners outside the survey.</p><p>What the data does show is that the contemporary beauty economy is moving beyond dramatic surgery and towards repeated maintenance. Japan’s large non-surgical market is particularly important because the future of cosmetic intervention may not be a completely different face. It may be a series of small treatments that maintain the impression that very little has changed.</p><h3>Different Societies Commercialise Beauty</h3><p>Japan and China illustrate how beauty can be industrialised without always appearing surgical. In Japan, maintenance is often gradual and subtle, which lowers the psychological barrier to intervention. A person who would reject major facial surgery may be comfortable with lasers, skin treatments, injectables or repeated rejuvenation procedures because these are presented as extensions of grooming rather than fundamental changes to identity.</p><p>China shows how quickly beauty pressure can operate at digital scale. Filters can smooth the skin, reshape the jaw, enlarge the eyes and change facial proportions before the user has consciously decided that anything is wrong. The altered face then becomes the standard against which the real face is compared. This is psychologically different from traditional celebrity comparison because people are no longer comparing themselves only with models and actresses. They are comparing themselves with an edited version of themselves.</p><p>Thailand and Vietnam represent another part of the beauty economy, where domestic demand exists alongside the growth of aesthetic medicine and international treatment. Thailand has established a strong reputation for medical tourism because of relative affordability, specialist experience and services designed for foreign patients. Vietnam’s reported figures were more surgery-heavy than Japan’s, although this may reflect the clinics and practitioners included in the survey rather than a national difference in insecurity.</p><p>Malaysia and Singapore sit within the same regional movement but represent different forms of access and presentation. Malaysia attracts patients from neighbouring countries and can offer treatment at lower costs than Singapore, while Singapore’s market is smaller, more expensive and strongly connected with regulation and medical reputation.</p><p>The relatively low raw number in Singapore does not prove that appearance matters less. Its small population, the possibility of receiving treatment overseas, higher prices and a preference for dermatology or less visible procedures may all affect the figures. My view is that beauty intervention in Singapore is often expected to remain understated, although this should be understood as an interpretation rather than an established national psychological fact.</p><p>Wealth may not reduce the desire to look younger or more attractive. It may simply allow intervention to become more private, gradual and difficult to detect.</p><p>Australia provides a useful contrast because it demonstrates how beauty demand can coexist with more explicit psychological safeguards. Australian guidelines require patients seeking cosmetic surgery to be assessed for suitability and screened for body dysmorphic disorder using a validated tool. The assessment must be conducted by the practitioner responsible for the surgery and cannot simply be delegated to sales staff or clinic employees.</p><p>India and the Philippines show that the social importance of beauty cannot be understood through procedure counts alone. India has a large treatment market, but its population is also enormous, which makes raw totals misleading. Its beauty culture also remains closely connected with colourism, class, marriageability, media representation and the persistent commercial suggestion that lighter skin creates greater social opportunity.</p><p>The Philippines has a highly visible beauty culture shaped by entertainment, pageants, social media and public presentation, despite reporting a far smaller number of procedures than Japan or India. A society can place enormous value on female beauty without expressing that value primarily through surgery. Makeup, skincare, clothing, dieting, photography, filters and pageant culture can create equally powerful expectations.</p><h3>Beauty Obsession</h3><p>The uncomfortable truth is that beauty carries social value because human beings repeatedly assign social value to it. Psychology describes an attractiveness halo effect in which people perceived as attractive are also more likely to be judged as intelligent, confident, emotionally stable, responsible, sociable and trustworthy. Large cross-cultural studies have found evidence of this pattern across many countries and regions.</p><p>This means women may not invest in beauty simply because they are vain or have been manipulated by advertising. They may understand, consciously or unconsciously, that attractive people can receive more attention, warmer treatment and more favourable assumptions. Once appearance affects employment, relationships, social inclusion and perceived competence, investing in beauty can become a rational response to an unfair environment.</p><p>Self-discrepancy also plays an important role. Psychological discomfort can emerge from the gap between the person someone believes she is and the person she believes she should be. Social media expands that gap by exposing women to an almost continuous stream of professionally styled, filtered and carefully selected faces. The ideal is not encountered occasionally. It is carried in the pocket and refreshed throughout the day.</p><p>Objectification adds another layer. Women can begin to experience themselves from the imagined perspective of an observer, monitoring how their bodies and faces may appear rather than simply living through them. Appearance becomes a permanent performance review in which the individual is both the employee and the evaluator.</p><p>Research continues to connect appearance-focused social media use with self-objectification, body-image concerns and interest in cosmetic procedures. A systematic review involving more than 13,000 participants found a consistent relationship among social-media exposure, body dissatisfaction and the consideration of cosmetic surgery.</p><p>Beauty filters make this cycle more personal because they do not simply show a more attractive stranger. They produce a more attractive version of the user. In one large study, the same individuals received higher ratings not only for attractiveness after beauty filters were applied, but also for traits including intelligence and trustworthiness.</p><p>The individual had not become more intelligent or trustworthy. The face had changed, and the observer’s interpretation of the person changed with it. That may be one of the clearest demonstrations of why beauty remains valuable and why people are prepared to spend so much time, money and emotional energy pursuing it.</p><h3>Surgery and Desire for Control</h3><p>It is easy to explain cosmetic surgery as insecurity, but that answer is too simplistic and often unfair. Some people have disliked a feature for many years, make an informed decision and experience genuine satisfaction after treatment. Others seek reconstructive or functional procedures following injury, illness, pregnancy, significant weight change or physical discomfort. Cosmetics can also be creative, enjoyable and closely connected with identity rather than shame.</p><p>For others, the appeal of treatment may come from control. Ageing, rejection, career insecurity and relationship problems can feel unpredictable, while a nose, eyelid, jawline or body area appears concrete and modifiable. The emotional problem is difficult to define, but the physical feature provides a visible target.</p><p>This does not mean the concern is imaginary. If society treats someone differently because of her appearance, the consequences are real even when the standard is unjust. However, surgery may be asked to solve problems that surgery was never designed to solve. A procedure can change a feature, but it cannot guarantee love, promotion, belonging, respect or emotional stability.</p><p>The best available evidence presents a more balanced picture than either the cosmetic industry or its critics often acknowledge. Some patients experience improved satisfaction with the treated area, increased confidence and positive psychosocial outcomes, while other outcomes remain uncertain and are influenced by expectations, mental health and the type of procedure.</p><p>A 2025 systematic review concluded that the strongest available evidence remained limited, particularly when examining broad and lasting improvements in mental health, quality of life and social functioning. This is why it is inaccurate to claim that cosmetic surgery never improves psychological well-being, but it is equally misleading to sell surgery as a reliable treatment for low self-worth.</p><h3>Dissatisfaction Becomes a Disorder</h3><p>Ordinary insecurity should not be confused with Body Dysmorphic Disorder (BDD). BDD involves a persistent and distressing preoccupation with perceived appearance flaws that may be minor or unnoticeable to others. People may repeatedly check mirrors, avoid photographs, compare themselves with others, conceal parts of their appearance or seek constant reassurance. Prabu</p><p>BDD appears at higher rates among people requesting aesthetic treatment than within the general population, which makes screening especially important. A technically successful procedure may fail to resolve the distress because the concern can remain, move to another feature or create demand for repeated correction.</p><p>The ethical responsibility of a clinician is therefore not limited to performing a technically safe procedure. It includes recognising when the request may be driven by a psychological condition and being prepared to delay or decline treatment.</p><p>A clinic that always says yes may appear customer-focused, but medicine is not ordinary retail. Sometimes responsible care requires refusing the sale.</p><h3>Beauty Helps and Harms</h3><p>Cosmetic and reconstructive procedures can provide legitimate benefits. Breast reduction may reduce physical discomfort, functional nasal surgery may improve breathing and eyelid surgery may improve vision when excess tissue obstructs the visual field. Aesthetic treatment can also increase satisfaction with a particular feature for appropriately selected patients who understand the limitations.</p><p>The health discussion becomes distorted when these functional benefits are combined with broad promises of happiness, confidence and a transformed life. Physical improvement and psychological recovery are not the same outcome, and one does not automatically produce the other.</p><p>Surgery carries the risks associated with anaesthesia, bleeding, infection, poor wound healing, scarring, nerve injury, asymmetry, blood clots and revision procedures. Non-surgical treatments also carry serious risks, despite being marketed as convenient or minimally invasive. The United States Food and Drug Administration warns that accidental injection of dermal filler into a blood vessel can result in tissue death, visual impairment, blindness or stroke. These complications are uncommon, but they can be permanent.</p><p>Travelling overseas for treatment does not automatically make a procedure unsafe. South Korea, Thailand, Malaysia, Singapore, India, Australia and other destinations contain highly qualified practitioners as well as providers whose standards may be less reliable. The important questions concern the qualifications of the surgeon, the accreditation of the facility, the quality of informed consent and the availability of aftercare.</p><p>The greatest medical-tourism risk can sometimes emerge after the procedure, when surgery has been performed in one healthcare system but complications must be treated in another. The original surgeon may no longer be easily accessible and the local doctor may not have complete records. Health authorities also warn that surgery and air travel can independently increase the risk of blood clots, while travelling during recovery can compound that risk.</p><p>A holiday package, airport transfer and luxury recovery room may make surgery feel like tourism, but the body does not recognise the marketing distinction and continues to experience surgery as surgery.</p><h3>Price Women Pay for Ageing</h3><p>The discussion about beauty is incomplete without ageing because the value assigned to female appearance is closely connected with youth. Men are often permitted to age into authority, experience and seriousness, while women are encouraged to age without appearing to do so.</p><p>This creates a form of beauty labour that includes not only money but also time, planning and emotional attention. Skincare routines, hair treatments, makeup, fitness, dental work, injectables and cosmetic procedures may all be treated as optional personal choices, yet the combined result can become an unspoken professional and social expectation.</p><p>A woman who participates may be criticised for being artificial, while a woman who does not participate may be criticised for looking tired, old or careless. The standard becomes impossible because women can be punished both for pursuing beauty and for appearing not to pursue it successfully enough.</p><p>What concerns me is not that women want to remain attractive. It is that attractiveness is increasingly presented as an individual responsibility and ageing as a failure of personal maintenance. Natural skin becomes unfinished, wrinkles become evidence of poor discipline and the face becomes a permanent improvement project.</p><h3>Ranking Reflection</h3><p>If we allow the commercial version of the golden ratio to determine beauty, South Korea comes first in my ranking, followed by Japan and China. Thailand, Australia and Vietnam occupy the middle, while the Philippines, India, Singapore and Malaysia follow.</p><p>However, the ranking does not prove that Korean women are naturally more beautiful than women elsewhere. It demonstrates that South Korea has built the most influential and internationally recognisable beauty system among the countries considered, and that this system aligns closely with the narrow facial proportions commonly rewarded by golden-ratio analysis.</p><p>That is a cultural achievement, a commercial achievement and possibly a public-health concern. It is not evidence of biological superiority.</p><p>The golden ratio appeals to us because it promises certainty. It suggests that beauty can be calculated, nationalities can be ranked and faces can be corrected. Psychology offers a less comfortable answer because it shows that attraction is shaped by familiarity, social reward, cultural exposure and the way attractive people are treated.</p><p>Women continue to invest in beauty because beauty can produce real advantages, even within a system that is unfair. Some cosmetic choices are informed, personally meaningful and genuinely beneficial, while others may emerge from discrimination, comparison, anxiety or the belief that an improved face will produce an improved life.</p><p>I do not believe the answer is to condemn cosmetics, surgery or the women who use them. The more important task is to question the conditions under which the choice is made, the expectations attached to it and the industries that profit from dissatisfaction while presenting themselves as the cure.</p><blockquote>Perhaps the most revealing question is no longer which country has the most beautiful women. It is why we remain so determined to rank women in the first place.</blockquote><h3>References</h3><p>International Society of Aesthetic Plastic Surgery. <em>ISAPS International Survey on Aesthetic and Cosmetic Procedures Performed in 2024</em>, released in June 2025.</p><p>Ministry of Health and Welfare, Republic of Korea. <em>Foreign Patients Surpass 2 Million in 2025, Positioning Korea as Asia’s Medical Tourism Hub</em>, released in April 2026.</p><p>Rossetti and colleagues. <em>The Golden Ratio, Dispelling the Myth</em>. Maxillofacial Plastic and Reconstructive Surgery.</p><p>Psychological Science Accelerator research collaboration. <em>Examining the Attractiveness Halo Effect Across Cultures</em>. Current Psychology.</p><p>Arab and colleagues. <em>Social Media Influence on Body Image and Cosmetic Surgery Considerations, A Systematic Review</em>.</p><p>Haddad and colleagues. <em>A Systematic Review on the Reinforcement Loop in Aesthetic Medicine</em>. Aesthetic Plastic Surgery.</p><p>Gulati and colleagues. <em>What Is Beautiful Is Still Good, The Attractiveness Halo Effect in the Era of Beauty Filters</em>.</p><p><em>The Psychosocial Outcomes Following Cosmetic Surgery Are Largely Unknown, A Systematic Review</em>. Journal of Plastic, Reconstructive and Aesthetic Surgery.</p><p>Medical Board of Australia. <em>Patient Assessment and Body Dysmorphic Disorder Screening Guidelines</em>.</p><p>United States Food and Drug Administration. <em>Dermal Fillers, Risks and Safety Information</em>.</p><p>United States Centers for Disease Control and Prevention. <em>Medical Tourism and Travel Following Surgery</em>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1720ca12202c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Two Fires of AI Adoption]]></title>
            <link>https://medium.com/@prabu.sangar/the-two-fires-of-ai-adoption-fc7ef12ca1c0?source=rss-ed44d3335f1d------2</link>
            <guid isPermaLink="false">https://medium.com/p/fc7ef12ca1c0</guid>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[ai-adoption]]></category>
            <category><![CDATA[layoffs]]></category>
            <category><![CDATA[digital-transformation]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Prabu Sangar]]></dc:creator>
            <pubDate>Sun, 05 Jul 2026 02:28:13 GMT</pubDate>
            <atom:updated>2026-07-05T02:28:13.333Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*ZznQYNTNqTSpLr-b" /><figcaption>Photo by <a href="https://unsplash.com/@galka_nz?utm_source=medium&amp;utm_medium=referral">Galina Nelyubova</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h3>General Adoption</h3><p>Every major technology wave begins with a promise. The steam engine promised scale. Electricity promised productivity. Computers promised accuracy. The internet promised access. Cloud computing promised flexibility. Artificial intelligence now promises something more intimate, intelligence on demand. It tells organisations that writing, coding, analysis, coaching, research, service, and decision support can happen faster than before.</p><p>In my opinion, the danger is not that companies are adopting AI. They should. The danger is that many companies are adopting AI during a period of financial pressure, investor impatience, and fear of being left behind. According to Computerworld, the technology sector’s 2023 layoffs came after a pandemic-era hiring surge, when digital demand rose sharply and later normalised. Computerworld cited Layoffs.fyi data showing about 262,682 tech layoffs across 1,186 companies in 2023, after 164,969 layoffs in 2022. That matters because AI did not arrive in a calm labour market. It arrived inside a workplace already being resized.</p><p>There are two fires burning inside organisations today. The first is the visible belief that AI will replace people and that smaller teams automatically mean better management. The second is quieter but equally dangerous, the rising cost of token usage. Every prompt, file upload, generated summary, agent workflow, model call, and repeated attempt consumes tokens, compute, credits, and money. Many companies govern headcount with discipline. In my view, they are not yet governing AI consumption with the same seriousness.</p><h3>Concept of Efficiency</h3><p>For more than a century, organisations have tried to make work measurable. Scientific management broke labour into tasks. Assembly lines standardised movement. Enterprise software standardised processes. Outsourcing moved work to lower-cost locations. Shared services centralised administration. Cloud computing turned infrastructure into consumption. Each wave carried the same business logic. When work can be measured, it can be optimised. When it can be optimised, waste can be reduced. When waste falls, margins improve.</p><p>I believe the logic is not wrong. Businesses must be efficient to survive. But somewhere along the way, efficiency became emotionally neutral. A company could remove layers and call it simplification. It could outsource work and call it globalisation. It could close offices and call it footprint optimisation. Today, it can reduce people and call it AI transformation. This is where I feel leaders must be more honest. A layoff may be financially necessary. A restructure may be strategically justified. A role may genuinely change because of technology. But when AI becomes the blanket explanation, the human story disappears.</p><p>Harvard Business Review has warned that companies may be laying off workers because of AI’s potential, not necessarily because of proven current capability, and that this can create cynicism, damage trust, and force companies into embarrassing rehiring later. That distinction matters. Cutting because a process has truly been redesigned is one thing. Cutting because a future tool might one day absorb the work is another.</p><h3>Imminent Cuts</h3><p>Microsoft, LinkedIn, and Google matter in this conversation because they shape corporate language. What Big Tech does often becomes the vocabulary of everyone else. Smaller companies watch how these firms explain restructuring, how investors respond, and how quickly the public conversation moves on. Over time, difficult workforce decisions become normalised. The language softens before the pain does.</p><p>According to Reuters reporting published by Inc., Microsoft said in July 2025 that it would cut nearly 4 percent of its workforce, around 9,000 workers, after a previous round affecting around 6,000 employees in May. The same report connected the cuts to Microsoft’s effort to rein in costs amid heavy AI infrastructure investment. LinkedIn, owned by Microsoft, announced 668 cuts in October 2023, according to TechCrunch, bringing its total cuts that year to 1,384. Google also announced about 12,000 role reductions in January 2023 through a message from Sundar Pichai, who said the company had hired for a different economic reality and needed to focus resources on priorities including AI.</p><p>My concern is not that these companies made cuts. Large organisations will always restructure. My concern is that AI is becoming the language that makes cuts sound inevitable, intelligent, and almost progressive. The market may hear discipline. Employees may hear something colder. They may hear that they must become part of the AI story or risk being written out of it.</p><h3>Replacement Story</h3><p>There is a difference between using AI to redesign work and using AI as a convenient explanation for decisions leaders already wanted to make. AI removes some tasks, changes many roles, creates new capabilities, and exposes old inefficiencies. But it does not automatically tell leaders what kind of organisation they should build. That responsibility remains human.</p><p>According to Challenger, Gray and Christmas, AI was the leading reason cited for U.S. job cuts in June 2026 for the fourth consecutive month. AI was cited in 14,029 cuts in June and 101,743 cuts through the first half of 2026. Yet the broader labour story is not only about disappearance. The World Economic Forum’s Future of Jobs Report 2025 projects 170 million new roles and 92 million displaced roles by 2030, resulting in a net increase of 78 million jobs. The Economist has also framed AI as more likely to reshape white-collar work than wipe it out, drawing on the longer history of technological revolutions.</p><p>My thought is this, when we step back from the headlines and look at the work itself, AI is not just a replacement story. It is a redesign story. If leaders treat it only as a replacement tool, they will cut faster than they rebuild. They will remove people before understanding the work those people actually hold. They will underestimate tacit knowledge, customer memory, cultural context, and judgment. A spreadsheet can show salary savings. It cannot show the cost of losing someone who remembers why the previous transformation failed.</p><p>Economist Daron Acemoglu has warned that AI may raise average productivity while also replacing workers or degrading job quality for those who remain. His argument is important because productivity does not automatically become shared prosperity. In my opinion, this is the point many leaders avoid. The question is not only whether AI makes work faster. The question is who benefits, who carries the risk, and whether the people who helped build the company are given a serious path into its next version.</p><h3>Trust Placement</h3><p>When leaders speak about AI transformation, they often use the language of systems. They talk about platforms, agents, operating models, productivity, workflow automation, and cost takeout. Employees hear a different language. They hear uncertainty. They hear relevance. They hear survival. They hear the silent question of whether they will still matter next year.</p><p>Simon Sinek’s writing on trust is especially relevant here. According to Sinek, organisations undermine trust through actions such as mass layoffs, false promises, and excessive control without autonomy. In my view, AI adoption will fail in many companies not because the tools are weak, but because the emotional foundation is weak. Employees will not experiment openly with a technology they believe is being used to quietly measure, rank, or replace them.</p><p>Harvard Business Review makes a similar point through the lens of psychological safety. It argues that AI integration often changes how teams work together, and that the real impact can show up as declining interpersonal trust, weaker coordination, and poorer decision-making. The recommendation is not to treat AI adoption only as a tech upgrade, but as a team development effort. That is the leadership work. It is slower than buying software, but without it, adoption becomes compliance rather than belief.</p><p>This matters because people do not learn well when they feel hunted. They protect themselves. They hide mistakes. They stop asking basic questions. They nod in meetings and disengage after. A company may see dashboards showing AI usage, but beneath the dashboard, employees may be wondering whether every efficiency they create will become evidence against their own job.</p><h3>Valuable Tokens</h3><p>The second fire is less emotional but just as serious. Most employees experience AI as a simple chat box. They type, receive an answer, and move on. It feels light. It feels almost free. But underneath that experience sits a metered economic model. Every prompt, uploaded file, long conversation, generated report, image, audio file, and agent action consumes tokens.</p><p>According to Deloitte, tokens are becoming the fundamental unit of AI economics, and leaders should govern AI as an economic system rather than manage it with outdated cost models. Deloitte also says organisations should use real-time monitoring, budget alerts, chargebacks, right-sized models, and FinOps practices to prevent runaway costs. The message is clear. AI spend is not a normal software licence. It behaves more like a living consumption engine.</p><p>The FinOps Foundation explains why this can become expensive in practice. It warns that real generative AI cost is shaped by operational details such as model choice, input and output differences, multimodal content, and context window growth. In long conversations, prior context may be repeatedly resent so the model can maintain continuity. What looks like one chat to a user can become a growing trail of token consumption to the company.</p><p>My concern is that many organisations apply strict discipline to people cost while allowing AI usage to grow with far less scrutiny. A manager may need multiple approvals to hire one analyst. The same team may run thousands of prompts, generate multiple versions of reports, activate agents, and consume model capacity without anyone asking whether the output created measurable value. That is not innovation. That is unmanaged consumption wearing the clothes of transformation.</p><h3>Scale</h3><p>Google’s own AI usage shows where enterprise AI is heading. At Google I/O 2026, Sundar Pichai said Google’s model APIs were processing about 19 billion tokens per minute and that more than 375 Google Cloud customers had each processed more than one trillion tokens in the previous 12 months. Most companies will never operate at Google’s scale, but the direction matters. AI is moving from a tool people open into a layer embedded across email, documents, CRM, HR systems, finance platforms, coding tools, customer service, analytics, and knowledge management.</p><p>Once AI becomes embedded everywhere, the question cannot simply be whether employees are using it. The better question is whether the usage is creating value. Microsoft’s own documentation for certain Microsoft 365 Copilot experiences describes cost dashboards, budgets, alerts, spending policies, prepaid credits, pay-as-you-go billing, and hard caps for usage-based billing through Copilot Credits. That is responsible design, but it also proves the larger point. If AI can be metered, it can run away. If it can run away, it must be governed.</p><p>Economist Enterprise has also highlighted the physical cost beneath AI, especially in data centres where cooling, power density, water use, and specialised infrastructure become critical as AI workloads rise. This matters because token usage is not just a line item in a software dashboard. It connects to compute, power, cooling, capital expenditure, and environmental pressure. The prompt may feel weightless, but the infrastructure behind it is not.</p><h3>Questioning Skills</h3><p>There is another risk that receives less attention than layoffs and token cost. It is deskilling. When employees use AI well, they can learn faster, produce better work, and move into higher-value activities. When they use it poorly, they may become dependent on outputs they cannot evaluate. The organisation may become faster while becoming intellectually weaker.</p><p>Harvard Business School research on generative AI and labour demand found that postings for occupations with many structured and repetitive tasks decreased after ChatGPT’s launch, while demand grew for roles involving analytical, technical, or creative work that AI can augment. The researchers argue that companies should view generative AI as an augmentation tool rather than merely a cost-cutting measure. I think this is the healthier path. The goal should not be to make humans smaller. The goal should be to move humans closer to judgment, empathy, creativity, and accountability.</p><p>This is also why cutting experienced people too quickly can be strategically dangerous. AI can read documents, but it does not remember why a policy was written. It can summarise customer feedback, but it does not know which relationship was saved by one difficult conversation. It can generate a restructuring plan, but it does not carry the scars of the last restructuring that failed. People do not only perform tasks. They carry context.</p><h3>Leadership Bubble</h3><p>The most dangerous framing in the AI conversation is the idea that the future is either people or AI. That is too simple. Machines should absorb work that is repetitive, rules-based, low-risk, and draining. Humans should remain central where judgment, ethics, relationship, ambiguity, and consequence matter. The challenge is not choosing one over the other. The challenge is designing the relationship properly.</p><p>In my opinion, leaders should stop asking how many roles AI can replace and start asking what work AI should absorb, what work humans should elevate, and what new value can be created because both exist. That question changes the emotional tone of transformation. It moves the conversation from removal to redesign. It forces leaders to think about training, trust, role evolution, governance, and long-term capability.</p><p>The mature organisations will not be the ones with the most AI tools. They will be the ones with the clearest AI discipline. They will know where AI should be used, where it should not be used, what it costs, what it saves, what risks it creates, and what human capability it strengthens. They will treat every AI business case as both a people decision and a cost decision.</p><h3>Our Call to Action</h3><p>Before we approve another AI tool, cut another role, or celebrate another efficiency target, we need to ask better questions. Are we replacing work that should genuinely be automated, or are we removing people before understanding the knowledge they carry? Are we measuring AI’s value, or merely its activity? Are we governing token usage with the same seriousness we apply to headcount? Are we building employee capability, or silently weakening it?</p><p>In my opinion, AI should not become an excuse for smaller leadership. It should become a demand for better leadership, where every automation decision has a human reason, every token has economic accountability, every employee has a path to relevance, and every organisation remembers that the future of work is not only about what machines can do, but about what humans still choose to protect.</p><h3>References</h3><ol><li>Computerworld, used for the pandemic-era hiring context and the scale of technology layoffs in 2022 and 2023.</li><li>Reuters via Inc., used for Microsoft’s July 2025 job cuts and the connection to cost control amid AI infrastructure investment.</li><li>TechCrunch, used for LinkedIn’s October 2023 layoffs and total 2023 reductions.</li><li>Google, used for Sundar Pichai’s January 2023 announcement of about 12,000 role reductions.</li><li>Challenger, Gray and Christmas, used for AI being cited as a leading reason for U.S. job cuts in June 2026.</li><li>World Economic Forum, used for projections on new and displaced roles by 2030.</li><li>The Economist, used for the argument that AI is more likely to reshape white-collar work than fully wipe it out.</li><li>MIT Economics and Daron Acemoglu, used for the warning that productivity gains may coexist with worker displacement and degraded job quality.</li><li>Simon Sinek, used for the discussion of trust, layoffs, and leadership credibility.</li><li>Harvard Business Review, used for the link between AI integration, psychological safety, and team trust.</li><li>Harvard Business Review, used for the warning that companies may cut because of AI’s potential rather than proven current capability.</li><li>Deloitte, used for AI token economics, FinOps, monitoring, chargebacks, and AI cost governance.</li><li>FinOps Foundation, used for hidden token costs, context window growth, and real generative AI unit economics.</li><li>Google Blog, used for token processing scale across Google’s AI services.</li><li>Microsoft Learn, used for Copilot Credits, usage-based billing, budgets, alerts, and hard caps.</li><li>Economist Enterprise, used for the physical infrastructure cost of AI data centres, including power, cooling, and water use.</li><li>Harvard Business School Working Knowledge, used for the distinction between automation-prone and augmentation-prone roles.</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fc7ef12ca1c0" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Professional Development and Artificial Intelligence]]></title>
            <link>https://medium.com/@prabu.sangar/professional-development-and-artificial-intelligence-f4062105243c?source=rss-ed44d3335f1d------2</link>
            <guid isPermaLink="false">https://medium.com/p/f4062105243c</guid>
            <category><![CDATA[future-of-work]]></category>
            <category><![CDATA[professional-development]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[education]]></category>
            <category><![CDATA[training]]></category>
            <dc:creator><![CDATA[Prabu Sangar]]></dc:creator>
            <pubDate>Tue, 30 Dec 2025 06:36:35 GMT</pubDate>
            <atom:updated>2025-12-30T06:36:35.702Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*kHsB02J4gx9s63M0" /><figcaption>Photo by <a href="https://unsplash.com/@lukejonesdesign?utm_source=medium&amp;utm_medium=referral">Luke Jones</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h3>Beginnings of Professional Development</h3><p>Professional development existed long before the advent of artificial intelligence, and its original purpose was neither <strong>innovation</strong> nor <strong>transformation</strong>, but <strong>stability</strong>. The lineage of professional development reflects how societies have historically valued <strong>human capital</strong>, <strong>expertise</strong>, and <strong>continuity of work</strong>. Far from being a contemporary corporate construct, professional development emerged from deeply embedded social and economic practices that balanced <strong>practical mastery</strong> with <strong>structured knowledge transmission</strong>.</p><p>Historically, professional development arose to ensure that individuals could perform <strong>defined roles</strong> consistently over time. In <strong>pre-industrial economies</strong>, this took the form of <strong>apprenticeship</strong> and <strong>on-the-job learning</strong>. Skills were transferred through <strong>direct observation</strong>, <strong>repetition</strong>, and <strong>correction</strong>, often across many years. <strong>Tacit knowledge</strong>, rather than formal documentation, was the dominant medium of learning. This period established a foundational principle that <strong>professional competence</strong> is not purely theoretical, but is formed through sustained practice embedded within real work contexts. The objective was not adaptability in the modern sense, but <strong>mastery within a stable occupational domain</strong>.</p><p>With <strong>industrialisation</strong>, this artisanal model reached its limits. As organisations expanded in size and complexity, professional development shifted from <strong>individual mentorship</strong> to <strong>standardised training programmes</strong>. By the early twentieth century, factories, civil services, and large enterprises relied on <strong>structured instruction</strong> to prepare workers for narrowly defined tasks. Training focused on <strong>procedural accuracy</strong>, <strong>compliance</strong>, and <strong>repeatability</strong>, reflecting the operational needs of highly systematised workplaces. The sociologist <strong>Max Weber</strong> captured this shift in his analysis of bureaucratic systems, noting that modern organisations depend on <strong>predictability and control</strong>, conditions that fundamentally shaped how learning and competence were defined.</p><p>By the <strong>mid-twentieth century</strong>, professional development had become <strong>institutionalised</strong> on a global scale. This formalisation accelerated after <strong>World War II</strong>, particularly in the United Kingdom and other industrialised economies, as reconstruction and technological advancement increased organisational complexity. It became evident that <strong>initial education alone was insufficient</strong> for a lifetime of work. Corporations invested heavily in <strong>role-based training programmes</strong>, governments introduced <strong>mandatory continuing education</strong> for regulated professions such as medicine, law, teaching, and accounting, and universities expanded <strong>executive and professional education</strong> offerings.</p><p>The underlying assumption during this period was that <strong>knowledge durability</strong> could be relied upon. A professional could reasonably expect core skills to remain relevant for many years. This belief was supported by empirical evidence from the late twentieth century. Studies from this period indicate that <strong>technical and professional skills often retained relevance for 10 to 20 years</strong>, with <strong>incremental updates</strong> rather than wholesale replacement. Professional development therefore functioned as <strong>periodic renewal</strong>, not continuous reinvention. <strong>Expertise accumulated linearly</strong>, and <strong>seniority</strong> served as a reliable proxy for competence and authority.</p><p>This era also gave rise to <strong>lifelong learning</strong> as a socio-economic necessity. However, lifelong learning during this period was framed as <strong>episodic rather than continuous</strong>. Learning occurred at defined intervals and remained largely <strong>separate from daily work</strong>.</p><blockquote>Management theorist <strong>Peter Drucker</strong> articulated this worldview succinctly when he stated, <strong>“The most valuable asset of a 21st century institution will be its knowledge workers and their productivity”.</strong></blockquote><p>At the time, productivity was assumed to be sustained through structured updates to relatively stable bodies of knowledge.</p><p>Even as <strong>digital technologies</strong> began to emerge in the late twentieth century, the foundational assumptions of professional development remained largely unchanged. Early online learning platforms <strong>replicated classroom-based instruction</strong> rather than reimagining learning itself. Corporate training initiatives <strong>digitised content delivery</strong> while preserving the belief that learning was an <strong>event distinct from work</strong>. Completion of courses, certifications earned, and hours logged became dominant indicators of professional growth.</p><p>As a result, contemporary professional development frameworks were not designed for an environment in which the <strong>pace of change outstrips institutional training cycles</strong>. They were built for a world in which <strong>skills aged slowly</strong>, <strong>roles were clearly bounded</strong>, and <strong>expertise could be stabilised through periodic intervention</strong>. The emergence of artificial intelligence has not created this misalignment, but it has <strong>exposed and intensified it</strong>. When the lifespan of skills compresses from decades to a few years, and in some cases to <strong>months</strong>, the historical foundations of professional development are no longer sufficient to meet real-world demands.</p><h3>The Traditional Models</h3><p>Artificial intelligence does not merely introduce <strong>new skills</strong>. It destabilises the foundational assumptions upon which <strong>professional development</strong> was historically built. Traditional professional development has long rested on three core premises and they are <strong>skills change gradually</strong>, <strong>roles remain recognisable</strong>, and <strong>learning can occur outside of work</strong>.</p><blockquote>Artificial intelligence undermines all three.</blockquote><p>This destabilisation of professional development assumptions has been examined extensively in recent scholarship on generative artificial intelligence and learning. In particular, <strong>Dr Samson Tan’s analysis of generative artificial intelligence highlights how learning is no longer separable from work</strong>, but is increasingly embedded within evolving systems of practice.</p><blockquote>As Tan observes, “Generative artificial intelligence does not simply accelerate learning; it fundamentally alters what it means to learn, who learning is for, and how responsibility for learning is distributed”.</blockquote><p>This insight is critical because it reframes professional development not as a discrete intervention, but as an ongoing process of <strong>sense-making, judgment, and adaptation</strong> within volatile environments. Rather than preparing individuals for stable roles, professional development must now equip them to operate ethically and competently amid continuous change, a shift that traditional models were never designed to support.</p><p>This shift has been acknowledged at the highest policy levels. The Managing Director of the World Economic Forum, Saadia Zahidi, observed that <strong>“We are not necessarily looking at a negative future in terms of jobs, but we are looking at a major shift in the set of skills within each job and the types of jobs that will exist in the future”.</strong></p><p>This statement is significant because it reframes disruption not as job loss, but as <strong>task and capability transformation</strong>.</p><p>Empirical evidence reinforces this view. Recent global workforce analyses project that <strong>35% to 45% of job tasks will change materially by 2030</strong>, not because roles disappear, but because machines increasingly absorb core components of those roles. In parallel, more than <strong>60% of the global workforce</strong> is expected to require <strong>significant reskilling or upskilling</strong> within the next five years. These figures signal a <strong>structural transformation</strong>, not a cyclical adjustment.</p><p>The principal challenge facing organisations is therefore not ignorance of change, but continued reliance on <strong>obsolete professional development mental models</strong>. Many organisations persist in treating professional development as a finite intervention, assuming that once new skills are acquired, stability will return.</p><blockquote>This assumption is flawed.</blockquote><p>The dominant organisational response has been to emphasise <strong>reskilling initiatives</strong>, encouraging employees to attend short courses, acquire new certifications, and learn emerging tools. These efforts operate on the belief that professional development can restore equilibrium once new competencies are obtained. In reality, <strong>artificial intelligence evolves faster than institutional learning cycles</strong>. Tools change within months, workflows are reconfigured rapidly, and by the time a formal training programme is completed, the operating context has often already shifted.</p><p>The impact of this mismatch is especially visible in <strong>knowledge work</strong>. Roles such as analyst, consultant, researcher, and manager were traditionally defined by the ability to <strong>produce insight</strong>. Today, artificial intelligence systems can generate reports, forecasts, analyses, and recommendations in seconds. As a result, professional value migrates away from content production toward <strong>judgment</strong>, <strong>interpretation</strong>, and <strong>accountability</strong>.</p><p>Yet conventional professional development remains largely <strong>output-oriented</strong>. It trains individuals how to execute tasks efficiently, but rarely equips them to determine <strong>whether a task should be executed at all</strong>, or how to interrogate machine-generated outputs. This represents a critical gap between what professional development delivers and what contemporary work demands.</p><p>A similar transformation is evident in <strong>education</strong>. Generative artificial intelligence can automate lesson planning, assessment creation, and personalised feedback. Empirical studies indicate that AI-augmented instruction can reduce educators’ administrative preparation time by <strong>30% to 40%</strong>. This shift frees teachers and educators to focus on higher-order competencies that artificial intelligence cannot replicate, including <strong>ethical reasoning</strong>, <strong>student or learner motivation</strong>, <strong>emotional development</strong>, and <strong>critical thinking</strong>. These competencies do not fit neatly into traditional training modules, underscoring the limitations of existing professional development frameworks.</p><p>Beyond skills and tasks, artificial intelligence also reshapes <strong>professional identity</strong>. When core competencies are automated, accumulated experience and seniority no longer guarantee professional standing. This disruption often manifests as <strong>quiet resistance</strong>. Practitioners may comply outwardly with new systems while privately questioning their relevance and long-term value. Existing professional development programmes tend to address functional capability gaps, but they rarely provide the <strong>psychological and identity support</strong> required to navigate this erosion of expertise.</p><p>This omission marks one of the most visible failures of traditional professional development. Many organisations still frame artificial intelligence adoption as a <strong>technical upgrade</strong>, when in reality it constitutes a profound <strong>cognitive and psychological transition</strong> for the workforce.</p><p>Measurement practices further compound the problem. Learning effectiveness is still commonly assessed through <strong>hours completed</strong>, <strong>courses attended</strong>, and <strong>certificates earned</strong>, despite mounting evidence that these indicators correlate weakly with adaptability or sound judgment. Surveys consistently show that <strong>less than 25% of formal training leads to sustained behavioural change</strong> in the workplace. Under conditions of rapid technological change, this already low effectiveness rate becomes increasingly untenable.</p><p>Taken together, these dynamics reveal a deep <strong>misalignment between traditional professional development paradigms and the volatility of the artificial intelligence era</strong>. Conventional models were designed for environments characterised by relative stability, predictable skill decay, and clearly bounded roles. Artificial intelligence introduces continuous, systemic instability.</p><p>The implication is clear. The solution is not to intensify existing training modalities, but to <strong>redefine professional development itself</strong>. A reimagined framework must prioritise <strong>learning agility</strong>, <strong>sense-making</strong>, <strong>ethical reasoning</strong>, and <strong>human judgment</strong> over narrow task proficiency.</p><blockquote>As organisational theorist Chris Argyris argued, <strong>“Learning that does not question underlying assumptions is merely adaptive, not transformative”.</strong> Professional development that remains adaptive alone will continue to lag behind reality.</blockquote><h3>Who We Learn From</h3><p>One of the most consequential questions in the age of artificial intelligence is not <strong>what</strong> professionals learn, but <strong>who</strong> they learn from. Historically, professional development assumed a clear hierarchy of knowledge transmission. Learning flowed from <strong>instructors</strong>, <strong>mentors</strong>, <strong>peers</strong>, and <strong>institutions</strong>. Technology functioned only as a medium of delivery, not as a source of authority.</p><blockquote>Artificial intelligence disrupts this structure by positioning itself as both <strong>tool and tutor</strong>.</blockquote><p>This shift is already visible in practice. Surveys of knowledge workers indicate that more than <strong>60%</strong> now consult artificial intelligence tools weekly to generate explanations, summaries, or recommendations before approaching colleagues or formal learning resources. In some organisations, artificial intelligence has become the first point of reference for unfamiliar tasks, surpassing internal documentation and human experts. This behaviour should not be interpreted as disengagement, but as a rational response to speed, accessibility, and efficiency.</p><p>Artificial intelligence offers <strong>immediacy</strong>. It responds without judgment. It is available at the point of need. Traditional professional development rarely offers these attributes. Courses must be scheduled, mentors may be unavailable, and organisational knowledge bases often lag behind real practice. In contrast, artificial intelligence meets professionals directly within their workflow.</p><p>However, this convenience introduces a fundamental limitation. Artificial intelligence can generate answers, but it cannot generate <strong>contextual wisdom</strong>. It lacks understanding of organisational history, power dynamics, ethical nuance, and lived consequence. When professionals rely exclusively on artificial intelligence for learning, they risk developing <strong>technical competence without judgment</strong>.</p><p>This distinction has been emphasised by leading thinkers in education and technology.</p><blockquote>Bill Gates has repeatedly argued that <strong>“Technology is just a tool. In terms of getting people working together and motivating them, the human element remains central”.</strong></blockquote><p>While Gates originally framed this in the context of education, the insight extends directly to professional development. Information can be automated. <strong>Meaning, accountability, and ethical interpretation cannot</strong>.</p><p>This is where <strong>blended learning</strong> becomes structurally essential rather than pedagogically optional. Blended learning in the context of artificial intelligence is not simply a mix of online and offline formats. It represents a deliberate division of cognitive labour. Artificial intelligence supports <strong>information retrieval</strong>, <strong>pattern generation</strong>, and <strong>cognitive offloading</strong>, while humans provide <strong>interpretation</strong>, <strong>challenge</strong>, and <strong>value-based judgment</strong>.</p><p>Empirical research supports this distinction. Studies consistently show that learning outcomes improve when automated instruction is combined with human facilitation, particularly in domains involving decision-making and behavioural change. Programmes relying solely on self-directed digital learning report <strong>significantly lower retention and transfer rates</strong>, often below <strong>40%</strong>, compared to blended approaches that incorporate discussion, coaching, and reflection.</p><p>Professionals may acquire information quickly through artificial intelligence, but they frequently struggle to <strong>internalise</strong> it. They know what to do, but not why it matters. They can execute recommendations, but cannot defend decisions when challenged. Human-led elements in professional development address this gap. <strong>Mentors provide perspective</strong>, <strong>peers introduce productive friction</strong>, and <strong>facilitators create space for disagreement and sense-making</strong>. These interactions are not inefficiencies. They are the mechanisms through which learning becomes durable and defensible.</p><p>Some organisations have begun responding intentionally to this shift by reintroducing <strong>human friction</strong> into artificial intelligence-enabled learning environments. Professionals are encouraged to use artificial intelligence for preparation and exploration, but final decisions must be articulated, justified, and defended in human forums. Artificial intelligence outputs are treated as inputs rather than conclusions. Evidence from early adopters suggests that this approach reduces error rates, improves decision quality, and strengthens collective accountability.</p><p>This position is not formed solely through abstract theory. It has also been shaped through direct scholarly collaboration with Dr Samson Tan, a leading voice in adult learning, workforce pedagogy, and education transformation in Singapore.</p><p>Dr Tan’s work spans public policy, higher education, and professional capability development, and consistently interrogates the assumption that technology alone can resolve learning and skills challenges.</p><blockquote>Working alongside him has reinforced a central insight of this paper, that while artificial intelligence can scale access to information, learning itself remains deeply human and relational.</blockquote><p>As Dr Tan has articulated in his public lectures and policy engagements, “Technology can support learning, but it cannot replace the human work of sense-making, judgment, and responsibility”. This collaboration has deepened my conviction that professional development must remain anchored in human expertise, dialogue, and ethical accountability, particularly as intelligent systems become more pervasive.</p><p>There is also a clear <strong>generational dimension</strong> to this transition. Younger professionals, particularly those entering the workforce after the pandemic, demonstrate higher comfort learning from artificial intelligence than from traditional authority figures. Yet survey data shows that they still place high value on <strong>mentorship</strong>, <strong>feedback</strong>, and <strong>recognition</strong>, provided these interactions are meaningful rather than hierarchical. This suggests that professional development is not losing relevance, but that <strong>the role of the human expert is evolving</strong>.</p><p>Senior professionals are no longer required to function as repositories of information. Their value now lies in <strong>framing problems</strong>, <strong>transmitting judgment</strong>, and <strong>modelling ethical responsibility</strong>. Artificial intelligence can handle information abundance. Humans remain responsible for interpretation.</p><p>Professional development therefore must evolve from one-directional knowledge transmission toward a <strong>dialogic learning model</strong>. In this model, artificial intelligence supports exploration and efficiency, while human relationships anchor learning in accountability, context, and shared meaning.</p><blockquote>As organisational scholar Donald Schön observed, <strong>“The practitioner’s reflection-in-action is what allows expertise to function under uncertainty”.</strong></blockquote><p>Artificial intelligence can accelerate action, but reflection remains a human obligation.</p><p>Blended learning is not a compromise between tradition and innovation. It is a recognition that learning, like work itself, has become a <strong>human and machine collaboration</strong>. Professional development that fails to acknowledge this reality risks irrelevance. Professional development that embraces it gains the opportunity to become deeper, more resilient, and more aligned with the realities of contemporary work.</p><h3>Exposure of Professional Development</h3><p>Professional development has not been weakened by artificial intelligence. It has been <strong>exposed</strong>. Artificial intelligence has revealed structural weaknesses that existed long before its arrival, but were previously concealed by slower rates of change and relative skill stability. Much of what passed for professional development was fundamentally <strong>transactional</strong>. It prioritised <strong>delivery over depth</strong>, <strong>efficiency over judgment</strong>, and <strong>completion over capability</strong>.</p><p>For decades, this model persisted because information was scarce and expertise aged slowly. Training programmes functioned adequately when knowledge could be transferred, retained, and applied over long periods. Artificial intelligence removes that protection. When machines can generate content, insights, and recommendations instantly, professional development that merely <strong>transfers information</strong> becomes redundant. This redundancy is not a failure of learning, but a failure of <strong>design</strong>.</p><p>Technology leaders have acknowledged this shift openly. Bill Gates has argued that within the next decade, advances in artificial intelligence will make “great medical advice and great tutoring” widely accessible and effectively free.</p><blockquote>This observation matters because it signals the <strong>commoditisation of knowledge transfer</strong>, which has historically been the core function of professional development.</blockquote><p>The prevailing organisational response has been to interpret this disruption as a <strong>skills deficit problem</strong>. Employers respond by accelerating reskilling initiatives, encouraging professionals to learn new tools, obtain certifications, and complete short courses. While these efforts are well intentioned, they rest on the assumption that professional development can restore stability once new competencies are acquired. Evidence suggests otherwise.</p><p>Global labour data indicates that more than <strong>60% of workers</strong> will require significant reskilling within the next five years, while the average lifespan of technical skills has fallen from <strong>20 years to less than 5 years</strong> in many digital domains. Under these conditions, professional development cannot function as a corrective intervention applied after disruption occurs. The environment changes faster than formal training cycles can respond.</p><p>Artificial intelligence also reshapes how professional value is defined. Professionals are no longer assessed primarily on what they can produce, but on what they can <strong>interpret</strong>, <strong>justify</strong>, and <strong>take responsibility for</strong>. Output alone is no longer sufficient. This represents a profound shift in professional identity. When machines perform analytical and generative tasks once associated with expertise, human contribution moves toward <strong>judgment</strong>, <strong>ethical reasoning</strong>, and <strong>accountability</strong>.</p><p>This transition is psychologically destabilising. Research on occupational identity shows that when core competencies are automated, professionals often experience anxiety, disengagement, and resistance, even when productivity improves. Seniority and accumulated experience no longer guarantee authority. Yet traditional professional development programmes rarely address this dimension. They focus on functional upskilling, while neglecting the <strong>identity, confidence, and meaning</strong> challenges introduced by artificial intelligence.</p><p>This neglect exposes one of the most significant limitations of conventional professional development. Many organisations treat artificial intelligence adoption as a <strong>technical upgrade</strong>, when it is in fact a <strong>cognitive and moral shift</strong>. Professionals are asked to trust systems they did not design, defend outputs they did not generate, and make decisions whose consequences they may not fully understand. Professional development that does not prepare individuals for this responsibility fails its core purpose.</p><p>Measurement practices further illustrate this exposure. Learning effectiveness is still commonly assessed through <strong>hours completed</strong>, <strong>courses attended</strong>, and <strong>certificates earned</strong>, despite evidence that these indicators correlate weakly with real-world performance. Multiple studies show that <strong>less than 25% of formal training</strong> leads to sustained behavioural change. In an artificial intelligence-driven environment, where decisions carry amplified consequences, this failure rate becomes unacceptable.</p><p>The implication is unavoidable. Professional development must be fundamentally redefined. It cannot compete with artificial intelligence on <strong>speed</strong>, <strong>scale</strong>, or <strong>information access</strong>. It must instead focus on what artificial intelligence cannot replicate. This includes <strong>sense-making under uncertainty</strong>, <strong>ethical ownership</strong>, <strong>learning agility</strong>, and <strong>human judgment</strong>.</p><blockquote>Organisational theorist Chris Argyris captured this distinction when he stated, “Learning that does not question underlying assumptions is merely adaptive, not transformative”.</blockquote><p>Professional development that remains adaptive alone will continue to lag behind reality. Transformative professional development, by contrast, equips individuals to examine assumptions, challenge machine outputs, and act responsibly in ambiguous conditions.</p><p>The exposure revealed by artificial intelligence is therefore not a threat to professional development, but an opportunity. If professional development continues to define success through content delivery and credential accumulation, it will become irrelevant. If it reorients toward developing <strong>judgment, accountability, and ethical clarity</strong>, it becomes indispensable.</p><p>The question is no longer whether professional development matters. The question is whether organisations, institutions, and leaders are prepared to allow it to become what this moment demands.</p><h3>Conclusion</h3><p>In my opinion, professional development has not been weakened by artificial intelligence, rather it has been <strong>revealed</strong> for what it truly is and, in many cases, for what it is not. Artificial intelligence has exposed long-standing assumptions that professional development was built upon, assumptions of skill stability, predictable career trajectories, and the durability of expertise. These assumptions were reasonable in an earlier economic era. They are no longer defensible.</p><p>Global workforce data makes this reality explicit. The World Economic Forum estimates that <strong>50% of employees worldwide will require reskilling by 2025</strong>, while the average half-life of technical skills has fallen to <strong>less than 5 years</strong> in many knowledge-intensive roles. Under such conditions, professional development cannot function as a periodic intervention applied after disruption occurs. The environment evolves faster than institutional learning systems can respond.</p><p>The central failure of traditional professional development is not a lack of effort or investment, but a misalignment of purpose. Too often, professional development has been designed to transfer information, standardise behaviour, and demonstrate compliance. Artificial intelligence renders these functions increasingly obsolete. Information is abundant. Execution can be automated.</p><blockquote>What remains scarce is <strong>judgment</strong>, <strong>ethical responsibility</strong>, and the capacity to act with clarity in uncertain conditions.</blockquote><p>This shift requires a fundamental redefinition of professional value. Professionals are no longer evaluated primarily on what they can produce, but on what they can <strong>interpret</strong>, <strong>justify</strong>, and <strong>stand behind</strong>. Artificial intelligence may generate outputs, but accountability remains human. When professionals are unable to explain or defend decisions beyond referencing a system recommendation, organisational trust begins to erode.</p><p>Professional development must therefore assume a new mandate. It must prepare individuals not only to use intelligent systems, but to <strong>question them</strong>, <strong>contextualise them</strong>, and <strong>take responsibility for their consequences</strong>. It must cultivate learning agility rather than static mastery, ethical reasoning rather than procedural compliance, and reflective judgment rather than task efficiency.</p><p>This transformation is not merely technical. It is deeply human.</p><blockquote>As philosopher Eric Hoffer observed, <strong>“In times of change, learners inherit the future, while the learned find themselves equipped to deal with a world that no longer exists”</strong>.</blockquote><p>Professional development that continues to serve the learned alone will fail. Professional development that serves the learner, the reflective practitioner, and the ethically accountable professional will endure.</p><p>The choice facing organisations is clear. Professional development can remain a functional utility focused on certification, hours, and delivery. Or it can evolve into a <strong>human-centred architecture</strong> that supports meaning, responsibility, and adaptability in an artificial intelligence-mediated world.</p><blockquote>The first path offers comfort. The second demands courage.</blockquote><p>That choice will determine not only the future of professional development, but the quality of leadership, decision-making, and trust that organisations are willing to accept in the age of artificial intelligence.</p><p>I would also like to acknowledge <strong>Dr Samson Tan</strong> for his collaboration and intellectual generosity throughout this journey. Engaging with his work and working alongside him on questions of adult learning, workforce pedagogy, and education transformation has been both grounding and clarifying. His view that technological progress must remain anchored in human judgment, ethical responsibility, and lived learning has deeply influenced how I have approached this paper. I am sincerely grateful for the opportunity to learn through that collaboration.</p><h3>References</h3><ol><li>Argyris, C. (1977). <em>Double Loop Learning in Organizations</em>. Harvard Business Review, 55(5), 115–125.</li><li>Bauman, Z. (2000). <em>Liquid Modernity</em>. Polity Press.</li><li>Brookings Institution. (2023). <em>Automation, AI, and the Future of Employment</em>. Brookings Policy Papers.</li><li>Coursera. (2023). <em>Global Skills Report</em>. Coursera Research.</li><li>Deloitte. (2023). <em>Global Human Capital Trends</em>. Deloitte Insights.</li><li>Drucker, P. F. (1999). <em>Knowledge-Worker Productivity: The Biggest Challenge</em>. California Management Review, 41(2), 79–94.</li><li>Gates, B. (2023–2024). Public remarks on artificial intelligence and the future of work and education.</li><li>Harvard Business Review. Tamayo, J., Doumi, L., Goel, S., Kovács-Ondrejkovic, O., &amp; Sadun, R. (2023). <em>Reskilling in the Age of AI</em>. Harvard Business Review.</li><li>Hoffer, E. (1973). <em>Reflections on the Human Condition</em>. Harper &amp; Row.</li><li>IBM Institute for Business Value. (2022). <em>The Enterprise Guide to Closing the Skills Gap</em>. IBM.</li><li>International Monetary Fund. (2023). <em>Gen-AI: Artificial Intelligence and the Future of Work</em>. IMF Staff Discussion Notes.</li><li>LinkedIn. (2023). <em>Workplace Learning Report</em>. LinkedIn Learning.</li><li>McKinsey Global Institute. (2017). <em>Jobs Lost, Jobs Gained: Workforce Transitions in a Time of Automation</em>. McKinsey &amp; Company.</li><li>Ministry of Education, Singapore. (2023). <em>Future-Ready Learners and Pedagogical Transformation Initiatives</em>. Government of Singapore.</li><li>Organisation for Economic Co-operation and Development. (2019). <em>OECD Employment Outlook: The Future of Work</em>. OECD Publishing.</li><li>Organisation for Economic Co-operation and Development. (2021). <em>Adult Learning and COVID-19: How Much Informal and Non-Formal Learning Are Workers Missing</em>. OECD Publishing.</li><li>Pew Research Center. (2023). <em>AI and the Future of Work: Public Perceptions and Workforce Anxiety</em>. Pew Research Center.</li><li>Schön, D. A. (1983). <em>The Reflective Practitioner: How Professionals Think in Action</em>. Basic Books.</li><li>SkillsFuture Singapore. (2023). <em>SkillsFuture Movement and Lifelong Learning Framework</em>. Government of Singapore.</li><li>Tan, S. (Dr Samson Tan). (2022–2024). Public lectures and policy contributions on adult learning, workforce pedagogy, and education transformation in Singapore.</li><li>Udemy Business. (2023). <em>Workplace Learning Trends Report</em>. Udemy.</li><li>UNESCO. (2022). <em>Artificial Intelligence in Education: Policy Guidance</em>. UNESCO Publishing.</li><li>UNESCO. (2023). <em>AI Readiness Assessment: Malaysia Country Profile</em>. UNESCO Global Observatory on AI Ethics.</li><li>Weber, M. (1978). <em>Economy and Society</em>. University of California Press. (Original work published 1922).</li><li>World Economic Forum. (2023). <em>The Future of Jobs Report 2023</em>. World Economic Forum, Geneva.</li><li>World Economic Forum. (2025). <em>The Future of Jobs Report 2025</em>. World Economic Forum, Geneva.</li><li><em>Tan, S. (2025). </em>The rise of generative artificial intelligence: Implications for work and learning<em>. In </em>Handbook of Artificial Intelligence in Higher Education<em> (p. 442).</em></li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f4062105243c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Connected Unconnectedly]]></title>
            <link>https://medium.com/@prabu.sangar/connected-unconnectedly-903c95d381dc?source=rss-ed44d3335f1d------2</link>
            <guid isPermaLink="false">https://medium.com/p/903c95d381dc</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[human-connection]]></category>
            <category><![CDATA[ethics]]></category>
            <category><![CDATA[psychology]]></category>
            <category><![CDATA[technology-and-society]]></category>
            <dc:creator><![CDATA[Prabu Sangar]]></dc:creator>
            <pubDate>Tue, 30 Dec 2025 01:32:10 GMT</pubDate>
            <atom:updated>2025-12-30T01:32:10.730Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*wpJqcqtveSnUOVXa" /><figcaption>Photo by <a href="https://unsplash.com/@rishabhdharmani?utm_source=medium&amp;utm_medium=referral">Rishabh Dharmani</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h3>Always Connected</h3><p>The implications of technology’s influence on human connection are most visible not in extreme cases, but in ordinary, repeated moments of daily life. I observe this most clearly in public spaces. When I walk into a MRT train carriage or board a SBSTrasit public bus, the scene is strikingly consistent. Rows of people sit in close physical proximity, yet almost every head is lowered toward a smartphone. Conversations are rare, not that it was prevalent prior to the smartphone era but still, it’s a daunting scene today. Eye contact is fleeting. Physical closeness no longer implies social presence, even among loved ones.</p><p>This is not an isolated impression, but a pattern reinforced by empirical evidence. Globally, adults now spend an average of <strong>6 to 8 hours per day</strong> interacting with digital devices, with smartphones accounting for the majority of this time. In Singapore, smartphone penetration exceeds <strong>90%</strong>, and daily usage ranks among the highest internationally. The smartphone has become an extension of the individual, mediating not only communication, but attention, emotional regulation, and identity formation.</p><p>What is particularly striking is how this constant connectivity reshapes expectations of availability. I notice this most acutely during periods that were once associated with disconnection. Even while on holiday, individuals including myself, remain tethered to work through e-mail, Microsoft Teams, and messaging platforms. Time away from the office no longer equates to cognitive rest. Instead, it becomes a state of partial engagement, where one is physically present in one location while mentally anchored elsewhere.</p><p>Psychologists describe this condition as <strong>continuous partial attention</strong>, a state characterised by perpetual alertness to incoming stimuli without sustained focus on any single interaction. While this mode of engagement can create an illusion of responsiveness or productivity, research associates it with increased stress, reduced empathy, and diminished capacity for deep interpersonal connection. From a human perspective, the cost is subtle but cumulative.</p><blockquote>People are present everywhere, yet fully present nowhere.</blockquote><p>Social media intensifies this dynamic further. Beyond communication, it introduces an expectation of visibility. Experiences are increasingly documented, shared, and evaluated in real time. I observe how moments that could have remained private or intimate are instead filtered through the question of how they will appear to others. This behaviour aligns with what psychologists describe as <strong>fear of missing out</strong>, an anxiety that others are living more meaningful or connected lives elsewhere. Ironically, the effort to signal presence can erode the authenticity of the experience itself.</p><p>This raises a fundamental question about the nature of connection in a digitally saturated environment. Are these platforms fostering genuine relationships, or are they substituting depth with frequency?</p><p>Evidence suggests the distinction matters. While digital platforms are effective at maintaining weak ties and facilitating information exchange, they are far less reliable at sustaining emotionally rich relationships. Studies consistently show that high levels of social media engagement correlate with increased loneliness, particularly among younger users. Connection becomes something that can be counted, but not necessarily felt.</p><p>Importantly, this is not a rejection of technology. Digital tools have enabled long-distance relationships, supported vulnerable communities, and provided connection where none previously existed.</p><blockquote>The issue lies in <strong>substitution rather than supplementation</strong>.</blockquote><p>When constant contact replaces meaningful presence, connection risks becoming transactional. When communication becomes effortless, it can also become disposable.</p><p>What emerges, over time, is a quiet erosion of relational depth. Qualities such as patience, attentiveness, and tolerance for discomfort, traits historically cultivated through sustained human interaction, are no longer structurally reinforced. Instead, they must be consciously chosen, often in resistance to technological norms.</p><p>These everyday observations, repeated across contexts and generations, point toward a deeper transformation. Human connection is no longer something that unfolds slowly through shared time and vulnerability. It is increasingly accessed on demand, mediated by devices, and shaped by algorithms. This shift sets the stage for profound generational consequences, particularly for those who have never experienced connection without constant digital mediation.</p><h3>Generational Shifts</h3><p>The everyday patterns of constant connectivity do not exist in isolation. Over time, they shape expectations, behaviours, and ultimately, how different generations understand what it means to be connected at all. The erosion of presence observed in public spaces and private moments does not affect all age groups equally.</p><p>Older generations tend to remember a time when connection required effort and intentionality. Relationships were constrained by geography and availability. Missed calls were ordinary. Silence did not automatically signal neglect. For these cohorts, technology entered an already formed relational framework, supplementing habits that were learned offline.</p><p>Younger generations, particularly those born after the mid-1990s, have never experienced such constraints.</p><blockquote>For them, constant connectivity is not an innovation but a baseline condition.</blockquote><p>Communication is continuous, asynchronous, and ambient. Silence is often interpreted as absence. Delay is read as disengagement. Relationships unfold simultaneously across physical and digital spaces, often without clear boundaries between the two.</p><p>Psychological research suggests that growing up in such an environment reshapes expectations of availability and emotional responsiveness. Younger individuals report greater discomfort with unmediated interaction, including phone calls and unscripted conversations, preferring text-based communication that allows control and revision. At the same time, surveys consistently show higher reported levels of loneliness and anxiety among these cohorts, despite higher volumes of digital interaction.</p><p>This contradiction matters. It suggests that increased communication does not necessarily produce emotional security. Instead, constant comparison, visibility pressures, and responsiveness expectations can undermine relational confidence. Connection becomes something to manage rather than something to inhabit.</p><p>These generational differences are not failures of character or resilience. They reflect a deeper shift in the conditions under which human connection is learned. Those conditions now extend directly into the world of work.</p><h3>Work and Intimacy</h3><p>The same technologies that shape generational expectations of connection have also transformed the structure of work. As digital tools dissolve temporal and spatial boundaries, work becomes one of the most powerful forces shaping emotional availability.</p><p>I observe this repeatedly in ordinary settings. People travel, yet remain fixed to their devices. Meals are interrupted by notifications. Conversations pause mid-sentence to respond to work messages. Hypocritical observation yet I am equally guilty of such behaviour.</p><blockquote>Presence becomes conditional, always subject to interruption.</blockquote><p>Research reinforces this observation. Studies show that a large proportion of professionals check work communications multiple times per day while on leave. The boundary between professional and personal identity has eroded. Individuals are rarely fully at work, but they are also rarely fully away from it.</p><p>From a psychological perspective, this persistent connectivity alters intimacy. Intimacy requires sustained attention, not merely proximity. When attention is fragmented, emotional attunement suffers. Partners feel unheard. Friends sense distraction. Relationships do not collapse dramatically; they thin gradually.</p><p>Artificial intelligence accelerates this process. As productivity expectations rise, individuals rely on automation to cope. Messages are drafted, responses are optimised, and communication becomes efficient but flattened. Emotional texture gives way to speed and convenience.</p><p>Over time, this efficiency-oriented mode of interaction reshapes relational norms. Conversations become transactional. Emotional labour is minimised. The space for vulnerability narrows. In such conditions, intimacy competes directly with productivity.</p><p>Increasingly, this shift is visible in how care itself is expressed. Artificial intelligence is now routinely used to draft birthday wishes, anniversary messages, and even condolence notes. Most recipients are aware that these messages may be generated or assisted by AI, and yet many seem content to receive them. The words arrive on time, the tone is appropriate, and the social obligation is fulfilled.</p><p>What is quietly displaced, however, is the act of reaching out in ways that once required effort and vulnerability. Picking up the phone. Pausing long enough to ask how someone is coping. Sitting with discomfort rather than optimising around it. As expressions of care become more efficient, the practices that once sustained sincerity begin to fade. This raises an uncomfortable question. In catering to new expectations of speed and convenience, are we also allowing older, more demanding forms of care to disappear. And if so, does meeting one generation’s comfort with mediated empathy risk alienating the very practices through which emotional depth was once learned.The cumulative effect is not exhaustion alone, but disconnection. When every moment is potentially productive, no moment is fully relational. This creates the conditions under which presence itself becomes fragile.</p><h3>Relearning Presence</h3><p>If constant connectivity reshapes how generations relate and how work colonises attention, the remaining question is whether presence can be reclaimed at all.</p><p>Relearning presence is not about rejecting technology. It is about restoring intentionality. Presence can no longer be assumed as a default condition, it must be actively chosen.</p><p>This begins with recognising that attention is finite. When attention is continually divided, relationships absorb the cost. This is not a moral argument but a practical one. Human connection requires sustained, undivided engagement to develop trust, empathy, and understanding.</p><p>I observe small but meaningful responses to this reality. Phones placed face down during meals. Notifications muted during conversations. Work messages delayed rather than answered immediately. These gestures may appear trivial, but they signal a reassertion of agency.</p><p>At a broader level, relearning presence requires a shift in how society values time. Moments of silence, rest, and unmediated interaction are often treated as inefficiencies. In reality, they are the conditions under which emotional resilience and relational depth emerge.</p><p>The deeper challenge lies in resisting the belief that visibility equals existence. Social media rewards constant signalling, the need to demonstrate engagement and relevance. Yet connection does not require broadcasting. It requires attention.</p><p>As artificial intelligence increasingly mediates communication, this distinction becomes more critical. AI can assist with coordination, efficiency, and expression. It cannot substitute for the shared vulnerability that arises when people are fully present with one another.</p><p>Relearning presence is therefore not nostalgic or regressive. It is adaptive. The future of human connection will not be determined by the sophistication of our tools, but by the choices we make about how we inhabit the moments they mediate. Presence, once taken for granted, now requires protection.</p><h3>Choosing Connection</h3><p>I notice the shift most clearly in ordinary moments. Sitting across from someone I care about, both of us physically present, yet attention repeatedly pulled elsewhere. The mobile phone vibrates or lights up. A message arrives. A response feels required. Nothing dramatic happens, but something subtle is lost. The moment does not deepen. It simply passes.</p><p>What is at stake in these moments is <strong>trust</strong>. Not trust in a dramatic sense, but the quiet assurance that one has another’s full attention, even briefly. Trust is built when people feel prioritised, not when they are merely reachable. When attention is consistently divided, trust does not disappear, but it weakens. Relationships become easier to maintain, yet harder to deepen.</p><p>This is where ethics enters the conversation, not as regulation or abstract principle, but as daily practice. Ethics is present in the decision to pause before responding, to finish a conversation without interruption, to allow silence without filling it. These are small choices, often invisible, yet they shape how connection is experienced over time.</p><p>Future generations will absorb these norms long before they can articulate them. They will learn what connection means by watching how adults relate, how often work intrudes into personal space, and whether being unavailable is treated as neglect or as a boundary. If constant connectivity becomes the default, presence may come to feel unfamiliar rather than natural.</p><p>Artificial intelligence will continue to advance. It will become more capable, more conversational, and more embedded in everyday interaction. But no system can replicate what happens when someone is fully present, without distraction or optimisation. Trust grows in those moments, and responsibility for them remains human.</p><p>The challenge is not to disconnect, but to decide what deserves attention. Being connected is effortless. Being present requires intention. The future of human connection will depend on whether presence remains something we actively practise, rather than something we quietly surrender.</p><h3>References</h3><ol><li>Cacioppo, J. T., &amp; Patrick, W. (2008). <em>Loneliness: Human Nature and the Need for Social Connection</em>. W. W. Norton &amp; Company.</li><li>Cigna &amp; Ipsos. (2018). <em>U.S. Loneliness Index: Survey of 20,000 Americans Examining Behaviors Driving Loneliness in the United States</em>. Cigna.</li><li>Dunbar, R. I. M. (1992). <em>Neocortex Size as a Constraint on Group Size in Primates</em>. <em>Journal of Human Evolution</em>, 22(6), 469–493.</li><li>Maslow, A. H. (1943). <em>A Theory of Human Motivation</em>. <em>Psychological Review</em>, 50(4), 370–396.</li><li>McLuhan, M. (1964). <em>Understanding Media: The Extensions of Man</em>. McGraw-Hill.</li><li>Putnam, R. D. (2000). <em>Bowling Alone: The Collapse and Revival of American Community</em>. Simon &amp; Schuster.</li><li>Reuters. (2025, November 25). <em>AI Romance Blooms as Japanese Woman Weds Virtual Partner of Her Dreams</em>. Reuters News Service.</li><li>Turkle, S. (2011). <em>Alone Together: Why We Expect More from Technology and Less from Each Other</em>. Basic Books.</li><li>United Kingdom Government. (2018). <em>Tackling Loneliness: Appointment of a Minister for Loneliness and Government Strategy Announcements</em>. UK Government.</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=903c95d381dc" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Future of Corporate Travel]]></title>
            <link>https://medium.com/@prabu.sangar/future-of-corporate-travel-3945418aff97?source=rss-ed44d3335f1d------2</link>
            <guid isPermaLink="false">https://medium.com/p/3945418aff97</guid>
            <category><![CDATA[corporate-travel]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[human-centered-design]]></category>
            <category><![CDATA[future-of-work]]></category>
            <category><![CDATA[business-transformation]]></category>
            <dc:creator><![CDATA[Prabu Sangar]]></dc:creator>
            <pubDate>Sun, 07 Dec 2025 07:28:17 GMT</pubDate>
            <atom:updated>2025-12-07T07:28:17.682Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Travel should feel intelligent, humane, and worth the journey, not a relic of forms, friction, and after-the-fact reporting.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*dkSt4qbHe2oVUFhR" /><figcaption>Photo by <a href="https://unsplash.com/@2094_photography?utm_source=medium&amp;utm_medium=referral">Rachael Ren</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h3>Abstract</h3><p>I have always been fascinated by how two distinct ideas can converge to reshape an industry. In this paper, I explore how <strong>artificial intelligence</strong> and <strong>human-centered design</strong> are redefining corporate travel. Travel is no longer a simple series of bookings and approvals, it is becoming a <strong>strategic</strong>, <strong>predictive</strong>, and <strong>empathetic ecosystem</strong>. Drawing lessons from consumer technology, library sciences, fintech, and logistics, I argue that AI can transform travel into something <strong>purpose-driven</strong>, where <strong>intent</strong>, <strong>well-being</strong>, and <strong>outcomes</strong> matter more than policy enforcement. In my opinion, the future will belong to platforms that <strong>think, explain, and care</strong>.</p><h3>The New Frontier of Travel</h3><p>Corporate travel has crossed a line it cannot uncross, like seeing something you cannot unsee. Budgets are tighter, expectations are higher, and the old mechanics of <strong>“search, book, expense”</strong> no longer pass for strategy. In my view, business travel is shifting from a <strong>transactional cost</strong> to a <strong>lever for revenue, relationships, and culture</strong>. When I hear conversations with program owners and travelers, I hear the same refrain in different words and this is <strong>do not just process my trip, help me make it count</strong>.</p><p>This is not simply a technology problem. It is a purpose problem. Travel is no longer about sending people from point A to point B. It is about asking whether that trip builds trust, drives retention, unlocks collaboration, or shapes long-term partnerships. In my opinion, we are moving from <strong>how much we travel</strong> to <strong>why we travel</strong> and <strong>what each journey achieves</strong>.</p><p>A quote that has stayed with me quite a bit and it reminds me of the mindset shift that is needed, <strong>“Do not find customers for your products, find products for your customers.”</strong> In travel, the parallel is clear. Do not force travelers into rigid workflows that suit the system. Build systems that respond intelligently to travelers, their context, and their intent.</p><p>Industry data reinforces this shift. According to the Global Business Travel Association’s 2025 report, global business travel spending reached nearly <strong>1.5 trillion USD</strong> in 2024, with 2025 expected to surpass pre-pandemic levels. However, the story is no longer about raw volume. Across Asia, particularly in <strong>China</strong>, business travel has not only recovered but accelerated past 2019 levels. Both <strong>Trip.com</strong> and <strong>FCM</strong> have cited demand exceeding pre-pandemic benchmarks for the first time since recovery began. The conversation is changing from “How quickly can we restore travel?” to “How do we make travel <strong>worth it</strong> for the business and the traveler?”</p><p>In my opinion, this is where <strong>intent</strong> and <strong>human experience</strong> come to the forefront. Travelers are no longer satisfied with being treated as line items in a cost report. They expect the same level of intelligence and care that they see in their personal lives. If a consumer app can learn their preferences, anticipate needs, and adjust in real time, a corporate platform should do at least as much, and it should do so while honoring <strong>governance</strong>, <strong>duty of care</strong>, <strong>sustainability</strong>, and <strong>financial control</strong>.</p><h3>Honoring What Came Before</h3><p>A colleague shared with me recently, it is very easy to criticize systems that came before today and it is important to understand how much we have managed to get across using the same systems. He mentioned that it is important to <strong>pay homage</strong> to the foundations of corporate travel before we ask them to evolve.</p><p>For decades, <strong>GDSs</strong>, <strong>TMCs</strong>, and early <strong>expense and ERP systems</strong> were nothing short of transformational. They brought order to chaos. They digitised processes that were entirely manual. They provided a common language for airlines, hotels, agencies, and corporations to operate at scale. They ensured that travelers could move around the world with a reasonable expectation of safety, reliability, and financial traceability.</p><p>These systems were built for a world of fax confirmations, paper tickets, offline approvals, and batch reconciliation. They were designed to solve the problems of that era, fragmented content, limited visibility, and the need to centralise control. In that context, they were not just fit for purpose, they were visionary.</p><p>In my opinion, we do ourselves a disservice if we pretend that everything traditional is broken. The truth is that those systems got us here. The issue is that they were architected for a different set of questions. They were built to answer <strong>“What was booked?”</strong>, <strong>“What was spent?”</strong>, and <strong>“Was it within policy?”</strong>. Today, we are asking <strong>“Was this trip necessary?”</strong>, <strong>“Did it achieve its purpose?”</strong>, <strong>“Did the traveler arrive well enough to perform?”</strong>, and <strong>“What did we learn from how this trip unfolded?”</strong></p><p>The world has moved from <strong>control first</strong> to <strong>insight and experience first</strong>. The foundations of the past remain important, but the expectations placed on them have grown beyond their original design. The new frontier is not about discarding everything that came before. It is about building on that foundation to create systems that can <strong>understand context, adapt in real time, and care about outcomes</strong>.</p><p>In my opinion, this is where the real transformation lies. Not in replacing travel, but in <strong>reframing travel as a strategic, human-centered capability</strong> that supports the organisation rather than simply serving it.</p><blockquote><strong>The greatest danger in times of turbulence is not the turbulence; it is to act with yesterday’s logic. — Peter Drucker</strong></blockquote><h3>Where Current Systems Fall Short</h3><p>For many years, corporate travel systems did what the industry needed them to do. They brought consistency, compliance, and predictability to a process that involved thousands of variables. In my opinion, the issue today is not that these systems failed. It is that expectations have evolved faster than the systems were designed to evolve. The models that once defined efficiency now feel <strong>rigid</strong>, <strong>fragmented</strong>, and <strong>disconnected from intent</strong>.</p><p>Most platforms still operate as <strong>isolated islands</strong>. Booking tools collect itineraries. Expense systems collect receipts. Reporting systems collect history. Each component is useful on its own, but they rarely speak a common language in real time. Travellers feel this immediately. They move between tools that do not remember their choices, do not anticipate their needs, and do not adjust to the purpose of the trip.</p><p>In my experience, this is the root of the adoption challenge. People bypass official platforms not because they reject policy, but because they crave <strong>speed</strong>, <strong>clarity</strong>, and <strong>empathy</strong>. Travelers want the fluidity of consumer apps, the precision of financial systems, and the reassurance of a helpful assistant. When those expectations are not met, they turn to workarounds that feel faster and more intuitive.</p><p>The design gaps are easy to see. Forms ask for details the system should already know. Drop-down menus forget past behaviour. Policies list prohibitions rather than guiding travelers toward the best option. Reports describe what happened but rarely shape what should happen next. In my opinion, this is not a technology failure. It is a design philosophy problem. We talk about measuring value, but the systems we use still measure <strong>spend</strong>, <strong>exceptions</strong>, and <strong>compliance</strong> far more than experience, outcome, or well-being.</p><p>Data reflects this shift in sentiment. One global travel company reported that nearly <strong>40 percent</strong> of employees bypassed corporate tools because they found them confusing or restrictive. The same study revealed that the frustration was not with policy itself. It was with the systems that enforced policy in ways that felt bureaucratic or disconnected from the traveler’s reality.</p><p>In my view, this is where the industry must pause and reflect. Legacy systems were not built to ignore people. They were built for an era when standardisation and cost control were paramount. They succeeded at that mission. What they cannot do is reflect the nuance of <strong>purpose-driven travel</strong>. They were never asked to understand <strong>why</strong> a trip mattered or whether the timing of a flight would impact a critical client conversation.</p><p>Today’s world demands something more coherent. Travelers want systems that earn compliance through usefulness, not enforcement. They want clear choices that respect time, cost, carbon, and well-being. They want platforms that understand context so they can spend less time wrestling with forms and more time preparing for the reason they are traveling in the first place.</p><p>The industry is beginning to confront these gaps. New entrants and incumbents are experimenting with more integrated systems. Yet the fragmentation remains because the underlying architecture was never intended to adapt in real time. In my opinion, the next evolution will require not incremental patches, but a shift toward <strong>experience-centered</strong>, <strong>intent-aware</strong>, and <strong>intelligent orchestration</strong> layers that bind booking, policy, payment, and reporting into a single story. One of my favourite line which was meant for chaos, not people is that systems should never demand obedience. They should earn trust.</p><h3>A Glimpse of What Good Looks Like</h3><p>Every industry needs a reference point. In corporate travel, one of the clearest examples of meaningful reinvention comes from <strong>PwC’s collaboration with KAYAK and Blockskye</strong>. In my opinion, the reason this case stands out is not the technology itself, but the mindset behind it. PwC did not try to modernise an old system. They chose to rebuild travel from <strong>first principles</strong>, asking what the experience should be rather than what the constraints were.</p><p>The result was an ecosystem designed as a <strong>single source of truth</strong> for every reservation and every payment. Distributed ledger technology ensured that all parties saw the same information at the same time. There were no hidden fees, no reconciliation surprises, and no outdated records that required manual clean up. Visibility became a shared asset rather than a fragmented afterthought.</p><p>This approach delivered more than theoretical improvements. In its first twenty months, the platform processed more than <strong>one million</strong> PwC bookings. Intermediary fees dropped by roughly <strong>ten percent</strong>. Support calls fell by more than <strong>twenty percent</strong>, and response times improved by nearly <strong>thirty percent</strong>. Employees no longer needed to file expense reports for covered segments because internal billing codes handled the reconciliation automatically. Compliance improved simply because the workflow made the right path the easiest path.</p><p>In my opinion, this case matters because it proves that the problems in corporate travel are not inevitable. They are the product of intentional design. When booking, policy, payment, and reporting operate on a shared foundation, the experience becomes <strong>faster</strong>, <strong>clearer</strong>, and <strong>more humane</strong>. Travelers get the flexibility they expect. Finance teams get the accuracy they need. Program owners get real time visibility without additional administrative work.</p><p>Examples like this are beginning to appear across the industry. <strong>Amadeus</strong> is investing heavily in new retailing and servicing models through NDC and its expanded platform strategy. <strong>Booking Holdings</strong>, <strong>Trip.com</strong>, and <strong>SAP Concur</strong> are embedding conversational assistants directly into their user flows. <strong>Capital One</strong> recently acquired parts of <strong>Hopper’s travel engineering talent</strong>, a signal that fintech players see the same inefficiencies and believe travel can be rebuilt with modern, adaptive architectures. These moves are not just corporate announcements. I like to call them “nudge-indicators” that the travel ecosystem is preparing for its next structural shift.</p><p>In my opinion, what unites these initiatives is not technology. It is the recognition that <strong>traveler experience is a design choice</strong>. When systems become more honest, more transparent, and more responsive, adoption follows naturally. Travelers stop fighting the system because the system finally begins to work for them. Trust is earned, not enforced.</p><p>The future does not require every platform to replicate PwC’s model. It does require a shift in mindset. Good travel systems should not merely process actions. They should explain trade offs, understand intent, and help people make decisions that balance cost, time, carbon, and well being. When travel is designed with this clarity, it becomes something more than a workflow. It becomes <strong>intelligent infrastructure</strong> that supports the business and respects the human being at the center. A quote close to my heart is, <strong>“Perfection is not attainable, but if we chase perfection, we can catch excellence.” — Vince Lombardi</strong></p><h3>From Itinerary to Intent</h3><p>I was thought by my teachers that the simplest way to find out about something is to ask a question. In the same context, the most profound shift in corporate travel begins with a simple question. Instead of asking <strong>“Where are you going?”</strong>, the system should ask <strong>“Why are you traveling?”</strong> The first question triggers logistics. The second question triggers purpose.</p><p>This distinction changes everything. A board meeting in Tokyo demands different constraints from a multi city client loop. A long haul project handover deserves recovery time. A high stakes negotiation may justify a refundable fare, a shorter connection, or airport lounge access. When systems treat all trips as identical, they flatten the complexity of human work into a rigid template. In my view, this is the core limitation of itinerary centric design. The map becomes the territory, even when the territory is far more nuanced.</p><p>Purpose shapes the journey. It determines whether a traveler needs to prioritise rest, cost, proximity, safety, or predictability. It determines whether the meeting outcome depends on arriving alert or simply arriving on time. In my experience, travelers already think this way intuitively. Systems should simply catch up.</p><p>I learned this idea from an unexpected place and as a librarian, I am proud to say that it was found in my favourite place, <strong>libraries</strong>. Older catalogues could not understand language, synonyms, or context. They required exact matches. Discovery layers changed that. They could interpret how people searched and return materials that fit the idea rather than the literal phrase. Corporate travel needs the same evolution. Travelers should be able to type in plain language such as <strong>“Refundable flight to Tokyo next week, morning departure, within policy, show greener options”</strong> and receive results that automatically translate intent into filters, constraints, and recommendations.</p><p>In my opinion, this is the heart of intelligent travel. A platform should not ask travelers to become experts in fare codes, booking classes, or policy rules. It should do that work quietly and intelligently, surfacing clear choices that respect cost, time, carbon, and preference. In many cases, the barrier is not data. It is design. We often expect people to navigate complexity that systems could carry for them.</p><p>This is where the principle of <strong>explainability</strong> emerges. A traveler should see not only the recommended option but the reason behind it. <strong>“Five minutes from your meeting, within budget, lower on carbon, and flexible in case of schedule changes.”</strong> Simplicity is the truest form of intelligence. If the system cannot explain the reasoning simply, it does not yet understand the traveler or the context.</p><p>In my opinion, the move from itinerary to intent is not about adding features. It is about restoring clarity. Travel becomes strategic when every choice connects to purpose. When a system understands intent, it can pre rank results that fit the moment, highlight trade offs transparently, and prevent mistakes before they happen.</p><p>This shift matters because it changes adoption, satisfaction, and outcomes. Travelers trust platforms that understand them. Managers trust recommendations they can audit. Finance teams trust processes with fewer exceptions. Purpose is the foundation that unifies these stakeholders into a coherent system instead of forcing them into competing priorities.</p><p>Good travel design should remove friction, not add it. It should recognise that every trip has a story behind it. The more accurately a system understands that story, the more intelligently it can support the traveler.</p><blockquote>“If you change the way you look at things, the things you look at change.” — Wayne Dyer</blockquote><h3>Search Smarter</h3><p>In the concentrated experience I have gained, I feel that the way we search in corporate travel reveals how much of the burden we still place on the traveler. Search today is often a test of patience. The traveler types in a location, adjusts multiple filters, rechecks policy documents, guesses whether a fare is refundable, and then tries to infer which options balance cost, convenience, and corporate expectations. It is a process that is logical from the system’s perspective but exhausting from the human perspective.</p><p>My thinking on this changed years ago when I spent time studying how <strong>libraries</strong> transitioned from traditional catalogues to modern discovery layers. Old systems demanded precision. You had to know the exact term, the exact spelling, the exact structure of the query. Anything less produced nothing. Discovery layers took the opposite approach. They understood intent, synonyms, context, and user behaviour. They brought the right material forward without punishing the user for not knowing the system’s language.</p><p>Corporate travel needs this shift urgently. A traveler should be able to speak in <strong>plain language</strong> and trust the system to translate it into the correct technical parameters. For example, a search like <strong>“Morning flight to Osaka next week, refundable, near policy limit, greener if possible”</strong> should automatically interpret and convert into fare families, time windows, corporate caps, environmental scores, and risk conditions. The system should handle the complexity quietly, leaving the traveler with <strong>clarity</strong>, not cognitive load.</p><p>In my opinion, this is not about making the interface prettier. It is about respecting people’s time. Travelers should not need to understand the internal codes of the industry. They should not need to compare fare classes or decode change restrictions. They should be able to rely on a platform that understands the nuances of booking logic, policy constraints, traveller history, and the purpose of the trip.</p><p>This is where <strong>semantic search</strong> and <strong>contextual filters</strong> should become the standard. Instead of asking the traveler to navigate ten menus, the system should infer what matters most based on prior behaviour, organisational patterns, and stated intent. If a traveler consistently chooses flights that allow preparation time before important meetings, the system should recognise that pattern. If the company prioritises sustainability on short haul routes, the platform should surface greener options proactively.</p><p>Explainability matters too. In my opinion, a smart search system should not only present options but articulate the reasoning behind them. <strong>“Chosen because it aligns with your travel history, is within policy, and reduces carbon by twenty percent.”</strong> When travelers understand the why, they trust the results more deeply and rely less on external sources that fragment the experience.</p><p>This shift is not hypothetical. Consumer platforms already do this. Streaming services know when you skip a song. E commerce platforms infer intent from browsing patterns. Even navigation apps adjust routes based on past preferences. These systems adapt because they listen. Corporate travel systems should do the same.</p><p>I believe that smarter search is not a luxury. It is a basic commodity or foundation. It is how we turn travel from a technical exercise into a thoughtful, purpose aligned experience. It is how we help travelers make better decisions without forcing them to decode the language of aviation, hospitality, and finance.</p><h3>Agentic AI</h3><p>In my opinion, <strong>Agentic AI</strong> is not just another feature in the corporate travel stack. It represents a shift in how systems behave, how decisions are made, and how humans interact with technology. For years, AI in travel has been synonymous with chatbots, recommendation engines, and automated alerts. These tools were helpful, but they were reactive. Agentic AI moves the industry from <strong>reaction to collaboration</strong>, from <strong>information to execution</strong>, and from <strong>static workflows to adaptive intelligence</strong>.</p><p>The simplest way to understand this shift is to look at what Agentic AI actually does. It does not simply answer questions. It <strong>takes action</strong>. It connects insight to execution without waiting for manual triggers. Instead of telling you about disruption, it starts rebooking. Instead of showing you a policy rule, it shapes the options around that rule. Instead of handing you a list of flights, it holds compliant options and prepares the travel shell for manager review.</p><p>Here is a common example. A calendar invite lands for a conference in Paris. An agentic system immediately begins its chain of reasoning. It validates visa status. It checks traveler preferences and loyalty patterns. It applies corporate policy in context. It finds itinerary options that respect purpose, budget, and sustainability. It holds the best choices. It drafts the expense shell. It monitors travel conditions. It prepares fallback options if disruptions occur. By the time the traveler wakes up, what once required hours of searching, comparing, and checking has been reduced to a simple decision.</p><p>In my opinion, this is what human centered automation looks like. The goal is not to reduce headcount. It is to remove the small frictions that accumulate during travel planning. It is to prevent the red eye into an eight a.m. meeting. It is to ensure recovery time after a long haul flight. It is to offer a gentle nudge toward a train when the flight is short and the environmental impact is high. This is care delivered at scale, not efficiency delivered at the expense of experience.</p><p>The broader technology landscape points in the same direction. <strong>Anthropic</strong>, <strong>OpenAI</strong>, and <strong>Google DeepMind</strong> have all emphasised the rise of agentic systems that operate autonomously within guardrails. <strong>Elon Musk</strong> described a future where mobile apps dissolve and re form dynamically, replaced by intelligent nodes that create temporary experiences based on user intent. Whether or not this vision arrives exactly as described, the principle behind it is already affecting travel. People increasingly expect digital services to exist wherever they work, whether that is <strong>Slack</strong>, <strong>Teams</strong>, mobile, or voice.</p><p>In my opinion, travel will follow this pattern. Instead of logging into a portal, a traveler will simply say <strong>“Plan my trip to Osaka for the supplier meeting”</strong> and the system will construct a temporary workspace with booking flows, approval logic, transport links, and reporting structures. When the trip ends, the workspace disappears. Travel becomes ambient, not a destination. It appears when needed and dissolves when the purpose is fulfilled.</p><p>This shift requires travel teams to rethink design. Booking, policy, payment, communication, and reporting must live as flexible services, not rigid modules. They must integrate well enough to be orchestrated by a single intelligence layer. They must be transparent enough to be audited. They must be empathetic enough to be trusted.</p><p>The travel industry is already moving in this direction. <strong>SAP Concur</strong>, <strong>Trip.com</strong>, and <strong>Booking Holdings</strong> are embedding conversational agents that handle approvals and rebookings directly inside digital collaboration tools. <strong>Hopper</strong> and <strong>Capital One Travel</strong> are pushing the boundaries of predictive pricing and automated servicing. These examples show that the line between app and assistant is steadily dissolving.</p><p>In my opinion, the future of travel is not a tool. It is a <strong>presence</strong>. It is an intelligent layer that follows the traveler across devices, channels, and contexts. It respects purpose. It understands trade offs. It acts when allowed. It explains when needed. It supports the traveler without demanding attention.</p><h3>Spotify Effect</h3><p>In my opinion, one of the most important lessons for corporate travel does not come from travel at all. It comes from how <strong>Spotify</strong> interprets behaviour. Spotify does not rely on what users say they like. It watches what they skip, what they replay, what they listen to at different times of day, and how their patterns shift over weeks and months. It learns from signals, not just inputs.</p><p>Corporate travel has the same opportunity. Travelers make decisions that reveal intent long before they articulate it. A repeated switch from a preferred hotel may signal poor location fit. A pattern of declining early morning flights may point to fatigue. A tendency to extend stays around key meetings may reflect the need for preparation time. In my opinion, these are not inconveniences. They are <strong>data stories</strong> waiting to be understood.</p><p>Modern platforms should surface these signals automatically. If travelers upgrade long haul economy <strong>38%</strong> of the time on a specific route, that pattern deserves attention. If taxi costs are consistently high in a particular city because the preferred hotel is far from the office, that is a fixable design flaw, not traveler behaviour. If policy exceptions spike at quarter end, the issue might be timing, not non compliance.</p><p>In my experience, companies want to do the right thing for travelers, but they lack the feedback loops to make decisions based on real behaviour. Traditional reporting tells us what happened. Behavioural analytics explain <strong>why</strong> it happened. This is the difference between <strong>data</strong> and <strong>intelligence</strong>.</p><p>Several industry studies point in the same direction. Flight Centre Travel Group’s 2025 Business Travel Outlook found that companies using feedback driven analytics improved traveler satisfaction by more than <strong>twenty five percent</strong> and reduced policy exceptions by nearly <strong>one third</strong>. Those results did not come from punitive rules. They came from listening.</p><p>This principle extends beyond travel. Fintech companies identify risk patterns before fraud occurs. Ride hailing services adjust supply based on predictive heat maps. Streaming platforms refine recommendations continuously. In each case, systems learn by <strong>studying deviations</strong>, not by punishing them.</p><p>In my opinion, corporate travel programs should view deviations as insights rather than errors. If a traveler consistently chooses a hotel outside the preferred list, the question should not be why they disobeyed. The question should be what the system failed to recognise about their needs. If an employee repeatedly shifts flights at the last minute, the pattern may reveal something about their role, their schedule, or the nature of their work.</p><p>The smartest travel programs treat these signals as design feedback. They adjust policies where needed. They refine preferred hotels based on actual convenience. They improve guidance rather than tighten restrictions. They create a system that evolves with the organisation instead of freezing it in rules.</p><p>In my opinion, this is how corporate travel becomes truly intelligent. It listens. It adapts. It learns from every interaction. The goal is not perfection. The goal is continuous improvement shaped by behaviour rather than manual committees.</p><p>Travel should feel personalised, not policed. When platforms understand patterns the way Spotify understands music, travel programs can finally become <strong>adaptive</strong>, <strong>responsive</strong>, and deeply <strong>human</strong>.</p><h3>Competitive Analysis</h3><p>In my opinion, the most interesting developments in corporate travel are coming from organisations that are willing to start with a blank page. Instead of trying to modernise legacy systems piece by piece, these rising players ask a more fundamental question, “<strong>What should travel feel like today, and what architecture allows that to happen?</strong></p><p>The answer almost always involves <strong>modularity</strong>, <strong>open APIs</strong>, <strong>mobile first experiences</strong>, and <strong>integrated travel and expense stacks</strong> that remove duplication and reduce friction. These companies treat travel not as a string of disconnected steps, but as a continuous flow of context and decision making.</p><p>Platforms like <strong>Amex GBT</strong>, <strong>Hopper</strong>, <strong>Capital One Travel</strong>, <strong>Trip.com</strong>, and <strong>SAP Concur</strong> are experimenting boldly. They are merging <strong>fintech tools</strong> like price freeze, predictive rebooking, and automated dispute management with <strong>travel intelligence</strong> that understands behavioural patterns and policy logic. Consumer travel has pushed the boundary for years with dynamic inventory, rapid content refresh, and personalised recommendations. The lesson is not to copy leisure travel, but to translate its intelligence into models that respect governance and enterprise scale.</p><p>Fintech players, especially, have spotted the inefficiencies in travel and are moving quickly. Capital One’s acquisition of key engineering and hotel technology assets from Hopper signals that modern fintech firms see corporate travel as a domain ripe for reinvention. In my opinion, this combination of financial insight, behavioural modelling, and real time decision automation will shape much of the next decade.</p><p>Rising players also think differently about <strong>architecture</strong>. They treat orchestration as the primary capability. Instead of focusing on standalone features, they design systems that connect <strong>air</strong>, <strong>hotel</strong>, <strong>rail</strong>, <strong>policy</strong>, <strong>payment</strong>, and <strong>reporting</strong> into a single rhythm. When a meeting moves, the flight should adjust. When a trip extends, expense data should update. When disruptions occur, rebooking should trigger automatically. This is the difference between <strong>aggregation</strong> and <strong>orchestration</strong>. One delivers choice. The other delivers experience.</p><p>In my experience, travellers do not want more tools. They want fewer tools that work together. Rising players understand this. They see travel not as a catalog to browse, but as a series of outcomes to support. Their systems are not just repositories of fares and rates. They are engines that interpret intent, anticipate needs, and minimise administrative work.</p><p>This is also why they invest heavily in <strong>semantic search</strong>, <strong>preference learning</strong>, <strong>cross channel presence</strong>, and <strong>agentic workflows</strong> that operate autonomously. These capabilities are not gimmicks. They are the foundation of seamless travel experiences that feel personalised and intuitive.</p><p>What is important is that rising players do not view travel as a static industry. They view it as a dynamic system that should evolve with organisational behaviour, global conditions, and traveler expectations. Their advantage lies not only in technology, but in mindset. They expect systems to adapt rather than remain fixed.</p><p>In my opinion, the organisations that succeed in the next wave of travel will share this mindset. They will move beyond the transactional model and embrace travel as <strong>intelligent infrastructure</strong>. The winners will not be those with the most content. They will be those with the most <strong>clarity</strong>, <strong>context</strong>, and <strong>coherence</strong>.</p><blockquote>“The future is already here. It is just not evenly distributed.” — William Gibson</blockquote><h3>AI Adoption and Human Expectation</h3><p>In my opinion, one of the most overlooked truths in corporate travel is that <strong>expectations have shifted faster than systems</strong>. People now live with AI every day. They use assistants at home. They rely on predictive suggestions from their phones. They expect personalised feeds, instant answers, and seamless transitions between devices. When they return to their corporate tools, the contrast is impossible to ignore.</p><p>This shift is not driven by hype. It is driven by habit. The more people experience intuitive technology in their personal lives, the less patient they become with rigid enterprise systems that behave like outdated rulebooks. They do not want to fight the interface. They want the system to understand their intent, offer clear explanations, and respond with empathy when things go wrong.</p><p>This is why <strong>explainability</strong> matters so much. An AI recommendation is only useful when the traveler understands <strong>why</strong> it was presented. A simple line such as <strong>“Recommended because it is five minutes from your meeting, within budget, and lower on carbon”</strong> builds trust. It reassures the user that the system has considered multiple dimensions. Without explanation, automation feels arbitrary. With explanation, it feels intelligent.</p><p>Tone also matters. When disruptions happen, the difference between <strong>“Your flight is cancelled”</strong> and <strong>“I have moved you to the next available flight, and lounge access is arranged while you wait”</strong> is the difference between stress and relief. In my view, AI should not only be accurate. It should be respectful. People do not want systems that feel mechanical. They want systems that feel aware.</p><p>Data supports this shift. Recent studies show that <strong>84%</strong> of travelers trust AI recommendations only when reasoning and data sources are made transparent. Trust, not sophistication, determines adoption. It is not enough for a platform to be powerful. It must be understandable and predictable.</p><p>This principle extends to travel management teams as well. Managers want AI to help them make better decisions without losing control. Finance teams want automation that reduces manual reconciliation but preserves auditability. Risk and duty of care teams want visibility that improves safety without intruding on privacy. In my opinion, AI succeeds only when it aligns with these human expectations, not when it overrides them.</p><p>The industry has already begun shifting in this direction. <strong>SAP Concur</strong>, <strong>Trip.com</strong>, <strong>Booking Holdings</strong>, and <strong>Amex GBT</strong> are embedding personalized AI assistants directly into booking flows, disruption workflows, and approvals. The goal is not to replace human support, but to reduce noise so that human agents can focus on high empathy situations. This is a redefinition of roles, not a reduction of them.</p><p>In my view, AI should restore judgment at scale, not eliminate it. Technology can flag risks, highlight patterns, and automate predictable tasks, but it cannot understand the deeper reason behind a trip or the human stakes attached to it. That is where people remain essential. AI can assist, but humans interpret meaning.</p><p>This is why I believe the future of AI in travel is not about automation alone. It is about <strong>alignment with human expectations</strong>. The best systems will be those that think clearly, explain transparently, and respond with care.</p><blockquote>“The purpose of computing is insight, not numbers.” — Richard Hamming</blockquote><h3>One Ecosystem</h3><p>In my opinion, one of the most persistent challenges in corporate travel comes from treating <strong>travel</strong>, <strong>expense</strong>, <strong>duty of care</strong>, and <strong>sustainability</strong> as separate systems. Each one captures a different part of the journey, yet none of them deliver the full picture on their own. When these systems operate independently, the traveler becomes the integration layer. They become the person who copies data, reconciles records, and justifies exceptions. It is inefficient for the organisation and exhausting for the individual.</p><p>A more intelligent model treats the entire workflow as <strong>one ecosystem</strong>. Booking generates policy context. Policy informs approval paths. Approvals shape expense data. Expense reveals patterns that influence future policy. Duty of care layers safety signals across every step. Sustainability metrics provide visibility into environmental impact. These components should not live in isolation. They should reinforce one another, creating a cycle of clarity rather than a maze of tasks.</p><p>In my experience, the most frustrating moments for travelers usually occur because the system asks them to repeat something it already knows. They enter the purpose of travel multiple times. They reattach receipts that should have been matched automatically. They review line items that should have been pre coded based on itinerary data. None of this is intentional. It is a consequence of fragmentation.</p><p>This is where integration matters. When travel and expense flows are unified, the entire experience changes. Expense lines can be <strong>auto tagged</strong> to trip purpose before the traveler even departs. Likely exceptions can be flagged during booking instead of surfacing weeks later during reconciliation. Cost and carbon can be shown side by side so that decisions reflect budget and environmental commitments simultaneously.</p><p>Managers benefit as well. Instead of reviewing static dashboards after the fact, they can ask <strong>live questions</strong> such as:</p><p>· Which trips next quarter might exceed budget?<br>· Which routes consistently cause fatigue?<br>· What policy adjustments could reduce exceptions without hurting experience?</p><p>This kind of real time visibility turns reporting into guidance rather than history.</p><p>The industry is already recognising the value of unified workflows. According to SAP Concur’s latest study, companies using integrated travel and expense processes saved an average of <strong>12</strong> <strong>hours per traveler per trip</strong> in administrative work. That time is not just a productivity gain. It is time returned to meaningful work, creativity, and preparation.</p><p>In my opinion, the test for any ecosystem is simple. It should result in <strong>less typing</strong>, <strong>fewer surprises</strong>, and <strong>better decisions</strong>. It should support outcomes instead of forcing employees to navigate unnecessary forms. When systems are connected, travel becomes more predictable, more transparent, and far more aligned with the organisation’s goals.</p><p>Travel should never feel like a paperwork burden. It should feel like an experience that the organisation supports end to end. Bringing booking, policy, expense, duty of care, and sustainability into a unified ecosystem is how we move from fragmented tasks to <strong>coherent journeys</strong>.</p><p><strong>“Everything is connected. We just fail to see it most of the time.” — Alan Watts</strong></p><h3>My View Point</h3><p>In my opinion, the next generation of corporate travel should be built on four principles that guide every decision, every interaction, and every design choice. These principles are not features. They are values. They reflect what matters most to travelers, to organisations, and to the industry as it evolves.</p><p>The first principle is <strong>proactivity</strong>. Travel should begin with purpose, not process. A system should help determine whether a trip is needed, shape the best version of the journey, and anticipate risks before they materialise. Proactivity means giving people time back, not taking it away.</p><p>The second principle is <strong>explainability</strong>. Trust grows when systems reveal their reasoning. A traveler is far more likely to accept a recommended option when the system clearly shows how it aligns with cost, convenience, safety, and carbon goals. In my opinion, a transparent explanation is more valuable than a sophisticated algorithm that cannot articulate its logic.</p><p>The third principle is <strong>composability</strong>. Organisations should not be forced to adopt every capability at once. They should be able to add modules such as discovery layers, agentic booking, carbon accounting, behavioural analytics, or advanced reporting at their own pace. Composability allows systems to evolve alongside organisational maturity rather than demand a wholesale replacement in one step.</p><p>The fourth principle is <strong>human centered thinking</strong>. Corporate travel ultimately exists to support people. A system should care about outcomes, experience, and well being. A rested traveler closes deals. A calm traveler builds relationships. A supported traveler carries culture forward. Cost matters, but context matters more.</p><p>In my view, the real opportunity in corporate travel is to <strong>restore judgment at scale</strong>. Technology can automate decisions, highlight patterns, and reduce noise, but it cannot yet understand the deeper purpose behind a journey. It cannot know which meeting carries emotional weight or which conversation could shift the direction of a partnership. That is where human capability remains essential.</p><p>This is why I believe the best travel programs of the future will not compete on the number of integrations or dashboards they offer. They will compete on <strong>how well they interpret context</strong>, <strong>how clearly they explain decisions</strong>, and <strong>how effectively they support the traveler</strong>. The system must think, but it must also care.</p><p>These principles are not theoretical. They can be implemented today. They can shape how platforms prioritise features, how organisations refine policies, and how travelers experience the journey from planning to reimbursement. When these principles guide design, travel becomes more than logistics. It becomes an <strong>advantage</strong>, not an argument.</p><p>In my opinion, the path forward is clear. Begin with intent. Remove friction with agentic intelligence. Learn from every skipped option and deviation. Explain the why behind every choice. Fold cost, carbon, risk, and comfort into a single coherent view. Above all, design for people.</p><h3>The 1 to 3 Year Outlook</h3><p>In my opinion, the next one to three years will be defined by <strong>practical transformation</strong>, not dramatic reinvention. The industry is moving through a phase where clarity, integration, and efficiency matter more than futuristic predictions. Organisations want concrete improvements that reduce administrative load, improve traveler satisfaction, and create reliable visibility for teams across finance, procurement, and safety.</p><p>One of the clearest trends is the rise of <strong>AI copilots</strong> embedded directly inside booking flows. These copilots check policy, assist with routing choices, surface trade offs, and prepare itineraries that respect purpose and budget. They reduce manual steps and allow travelers to make informed decisions without reading long policy documents or switching between tools. Platforms like <strong>SAP Concur</strong>, <strong>Amex GBT</strong>, and <strong>Trip.com</strong> are already investing heavily in these capabilities.</p><p>Policies themselves will become more <strong>context sensitive</strong>. Instead of fixed rules, they will adjust based on purpose, trip length, traveler role, and sustainability targets. A short client meeting may justify an economy fare, while a long haul negotiation may require flexibility and recovery time. In my opinion, dynamic policy will become one of the defining features of modern travel governance.</p><p>Expense systems will also shift toward deeper integration with travel booking. When bookings, corporate cards, expense line items, and approvals sync in real time, reconciliation becomes almost invisible. Errors drop. Exceptions reduce. Finance teams gain clearer oversight without increasing administrative pressure on employees.</p><p>Reporting will evolve from traditional dashboards to <strong>live observability layers</strong>. These layers highlight anomalies as they occur, not weeks later. A trip that exceeds budget will trigger early alerts. A pattern of fatigue across certain routes will be flagged for review. Managers will gain the ability to ask natural language questions and receive immediate insights. In my experience, this shift turns reporting from a look back exercise into a decision making tool.</p><p>Industry data supports these shifts. According to Flight Centre Travel Group’s 2025 Business Travel Report, <strong>62%</strong> of companies plan to expand AI enabled expense automation within the next two years. More than <strong>half</strong> of travel managers expect AI to handle routine traveler support. This is not about replacing human agents. It is about freeing experts to focus on high empathy, high complexity cases where human judgment matters most.</p><p>Travelers will feel the difference in small but meaningful ways. Interfaces will predict intent. Policies will flex around purpose. Approvals will flow with fewer interruptions. Systems will explain their reasoning rather than enforce rules blindly. In my opinion, the winners in this phase will be the organisations that view AI as a collaborator rather than a monitor.</p><p>This period of practical transformation is essential. It builds the foundation for what comes next. It strengthens trust between travelers and systems. It reduces friction at scale. It brings coherence to workflows that have historically felt fragmented. It prepares organisations for a future where travel becomes more adaptive, more predictive, and more aligned with human experience.</p><h3>The 3 to 5 Year Horizon</h3><p>In my opinion, the three to five year horizon will be defined by <strong>visionary reinvention rather than incremental refinement</strong>. The boundaries between applications, devices, roles, and workflows will begin to dissolve. What we consider a travel platform today may not resemble what exists in the next decade.</p><p>A major influence on this shift is the growing concept of the <strong>AI node</strong>. Elon Musk described a future where mobile apps disappear and are replaced by dynamic, intent driven applications that form and dissolve based on user needs. Whether or not this vision arrives exactly as imagined, the underlying logic is already shaping technological expectations. People want systems that come to them, wherever they work, rather than demand that they log into static portals.</p><p>In corporate travel, this means that travelers may no longer open booking tools at all. They will simply state intent. They might say <strong>“Book my trip to Singapore for the supplier summit and follow policy”</strong> and the system will assemble a temporary workspace that includes routing choices, compliance checks, itinerary construction, risk assessments, and expense preparation. When the trip ends, the workspace closes. Travel becomes ambient, not a task. It appears when needed and disappears when it is no longer relevant.</p><p>This shift will also redefine travel roles. Routine tasks such as fare comparison, policy enforcement, disruption handling, and expense validation will be handled by <strong>agentic systems</strong> that operate autonomously within guardrails. Human capability will move upward into roles that require interpretation, empathy, negotiation, and strategic guidance. The travel consultant of today may evolve into roles such as <strong>Travel Intelligence Specialist</strong>, <strong>Corporate Mobility Strategist</strong>, or <strong>AI Performance Curator</strong>.</p><p>Industry data already hints at this direction. According to the FCM Global Report 2025, <strong>forty percent</strong> of TMCs are retraining staff for analytical and advisory work rather than transactional operations. This is not the removal of the human touch. It is the elevation of human contribution. Machines handle the mechanics. People shape meaning.</p><p>In my opinion, travel will also become more unified across channels. Booking, expense, approvals, communication, and reporting will operate as <strong>one adaptive layer</strong>. Travelers will interact through voice, mobile, collaboration apps, or even cars and wearable devices. Organisations will design travel ecosystems that integrate directly with procurement, finance, HR, and risk systems so that travel becomes a seamless part of how work gets done, not a separate process.</p><p>Sustainability will also become far more embedded. Carbon metrics will be visible at every decision point. Systems will automatically surface lower impact options when they align with purpose and policy. Environmental goals will live inside the workflow, not as external reporting requirements.</p><p>By this point, the industry will look very different from today. Systems will sense intent rather than react to inputs. Workflows will assemble themselves automatically. Predictions will guide decisions long before issues arise. Travel will feel more fluid, more intuitive, and far more aligned with outcomes and well being.</p><p>In my view, this shift represents the most significant redefinition of corporate travel in decades. It is not driven solely by technology. It is driven by new expectations of experience, purpose, transparency, and care. The systems that succeed will be those that interpret context as well as they process transactions.</p><p><strong>“We shape our tools, and thereafter our tools shape us.” — Marshall McLuhan</strong></p><h3>Product Ideas I Borrowed</h3><p>In my opinion, some of the most powerful ideas for the future of corporate travel come from outside the travel industry. These ideas remind me that inspiration does not need to be invented from scratch. It can be borrowed, translated, and applied in thoughtful ways that respect the complexity of travel while elevating the experience of the traveler.</p><p>One of my long held inspirations comes from <strong>libraries</strong>. Discovery layers changed everything in that world. A CTO can type <strong>“Meet supplier in Osaka next week”</strong> and the system should return filtered, intelligent, policy aware options the same way a well designed discovery engine returns relevant research materials. This model removes guesswork. It respects context. It interprets language with nuance.</p><p>From <strong>fintech</strong>, I borrow the idea of <strong>policy elasticity engines</strong>. Financial platforms examine exceptions, risk patterns, and behavioural signals to determine when a rigid rule should flex. Travel should do the same. When a deviation promises higher business value, the system should surface it transparently and capture justification automatically. This transforms rules from static constraints into dynamic, purpose aligned guidance.</p><p>From <strong>logistics</strong>, I take <strong>fulfillment intelligence</strong>. Global supply chains account for storms, labor strikes, customs patterns, and risk indicators long before they cause disruption. Travel systems should do the same. Visa lead times, geopolitical changes, or predictable disruption patterns should inform search results before the traveler ever sees them. In my experience, a proactive warning is worth more than a reactive apology.</p><p>From <strong>e commerce</strong>, I borrow recommendation engines that learn from more than clicks. Cancellations, rebookings, skipped options, travel purpose, and satisfaction scores should feed back into the system to refine future recommendations. Travel should learn from behaviour continuously, not occasionally.</p><p>From <strong>gaming</strong>, I take inspiration from <strong>adaptive feedback loops</strong>. Good games reward players for intelligent choices. Travel systems could reward sustainable selections, well planned itineraries, and choices that reduce exceptions. It is not about gamification for fun. It is about reinforcing behaviour that benefits the traveler and the organisation.</p><p>From <strong>consumer AI ecosystems</strong>, I borrow the idea of <strong>ambient support</strong>. The system should appear wherever the traveler already works. Slack, Teams, email, mobile, and voice interfaces should all serve as access points. Travel should not require the traveler to change channels. It should adapt to the environment.</p><p>These ideas have one thing in common. They shift travel away from rigid workflows and toward intelligent systems that continuously learn and refine. In my opinion, it is time to stop building portals and start building <strong>travel intelligence engines</strong>. Engines that listen. Engines that anticipate. Engines that simplify rather than overwhelm.</p><p>When travel systems borrow wisely from other industries, they become more than administrative tools. They become partners in how organisations move, connect, and grow.</p><p><strong>“The mind that opens to a new idea never returns to its original size.” — Albert Einstein</strong></p><h3>Practical Transformation</h3><p>In my opinion, the immediate future of corporate travel will be shaped by <strong>refinement rather than revolution</strong>. Organisations are not looking for dramatic upheaval. They want systems that work better, faster, and with less friction. They want meaningful improvements that reduce administrative burden, support decision making, and create a more predictable experience for travelers.</p><p>One of the clearest developments is the continued rise of <strong>AI copilots</strong> inside travel workflows. These assistants do more than surface options. They help interpret purpose, anticipate pain points, and streamline hours of repetitive tasks into a few guided steps. Booking flows will feel less like filling out a form and more like having a conversation that results in a set of informed choices.</p><p>Expense reconciliation will become increasingly <strong>predictive and automated</strong>. Instead of requiring employees to upload receipts, classify expenses, and correct mismatched entries, systems will auto match transactions, pre code cost centers, and raise exceptions before the traveler submits anything. The most valuable feature in this space will not be automation alone. It will be the removal of friction that historically exhausted busy professionals.</p><p>Policies will also become <strong>more flexible</strong>, adjusting based on travel purpose, sustainability commitments, and traveler context. In my experience, rigid rules do not make programs stronger. Contextually adaptive rules do. They reduce exceptions by understanding why travelers make certain choices in the first place.</p><p>This period will also bring improvements in <strong>reporting</strong>, shifting from static dashboards to tools that provide conversational insight. Managers will be able to ask plain language questions such as <strong>“Which departments exceeded budget this quarter?”</strong> or <strong>“Which routes create the most fatigue based on booking patterns?”</strong> and receive clear, actionable answers. Reporting will stop being a historical archive and instead become an intelligent guide for future decisions.</p><p>Systems like <strong>FCM</strong>, <strong>SAP Concur</strong>, and <strong>Trip.com</strong> are clearly moving in this direction, strengthening their integration layers and exploring partnerships that bring transparency and efficiency into a single workflow. This mirrors a wider trend in the industry where automation and data intelligence no longer sit at the edges of travel but at its core.</p><p>Trust will remain essential. Data privacy, explainability, and empathy will determine which systems succeed. In my opinion, automation should feel helpful, not intrusive. It should remove anxiety, not create it. The winning platforms in this period will be the ones that make each digital interaction feel natural and respectful.</p><p>The next one to three years will not be defined by futuristic predictions. They will be defined by <strong>steady improvement</strong>, <strong>continuous simplification</strong>, and <strong>smarter orchestration</strong> of the tools people already rely on. This is the groundwork that prepares the industry for the more transformative shifts that lie ahead.</p><p><strong>“Great things are not done by impulse, but by a series of small things brought together.” — Vincent van Gogh</strong></p><h3>Big Dreams</h3><p>In my opinion, the long term vision for corporate travel is far more ambitious than the incremental improvements we see today. The real transformation will occur when travel stops behaving like a set of tools and begins functioning as a <strong>fluid, intelligent ecosystem</strong> that forms and dissolves around the traveler.</p><p>This is where Elon Musk’s concept of the <strong>AI node</strong> becomes strikingly relevant. He imagines a future where applications do not live behind icons. Instead, dynamic micro experiences appear when intent is expressed and disappear once the task is complete. Whether this future arrives exactly as described is less important than the principle behind it. People increasingly expect technology to come to them rather than force them into predefined workflows.</p><p>For travel, this means a world where a traveler no longer visits portals or dashboards. They simply say <strong>“Book a trip to Malaysia next Tuesday for the regional review”</strong> and an AI node constructs everything needed. It builds the booking path, applies policy, integrates transport, prepares reporting links, and evaluates risks in real time. When the trip concludes, the entire workspace dissolves. Nothing remains except the outcomes and the insight gained.</p><p>In my experience, this shift will also redefine how organisations view the role of the travel consultant. Instead of spending time on repetitive servicing tasks such as rechecking fares, monitoring disruptions, or completing approvals, consultants will become <strong>strategic interpreters of intelligence</strong>. They will monitor AI behaviours, guide program optimisation, advise on sustainability impacts, and support complex, high stakes travel scenarios. Roles will evolve, not disappear.</p><p>Travel platforms will also function more like <strong>ecosystems of trust</strong>. AI will handle the logic. Humans will handle the meaning. Platforms will anticipate the traveler’s needs before the traveler expresses them, creating a journey that feels less like process and more like partnership. The system will learn from patterns, adapt to context, and align with the broader goals of the organisation.</p><p>Sustainability will be woven throughout this new architecture. Carbon visibility will sit beside cost visibility. Systems will detect when a greener alternative exists and make that option feel as natural as any other. In my opinion, sustainable design will no longer be a corporate initiative. It will be a default state.</p><p>By 2030, corporate travel may feel almost invisible, operating quietly in the background as a component of how work gets done. It will no longer be measured only by cost per trip. It will be evaluated by <strong>value per journey</strong>, impact on well being, contribution to strategic alignment, and its ability to support human connections that drive business outcomes.</p><p>In my view, these dreams are not distant fantasies. They are trajectories. Every improvement in AI autonomy, every enhancement in interoperability, and every shift toward intent driven experiences brings the industry one step closer to this future.</p><p><strong>“The future belongs to those who can imagine it, design it, and execute it.” — Mohamed bin Zayed</strong></p><h3>Reporting</h3><p>In my opinion, reporting has long been the most underleveraged component of corporate travel. For years, organisations relied on historical dashboards that summarised spend, exceptions, and compliance after the fact. These reports were accurate, but they rarely changed behaviour. They provided information but not insight. They documented what happened but did not influence what should happen next.</p><p>This is no longer enough. Executives today want <strong>observability</strong>, not archives. They want to understand <strong>patterns in motion</strong>, not just patterns in retrospect. They want real time deviation alerts, predictive forecasts, and summaries that read like decision briefs rather than spreadsheets. Reporting should not make leaders work to interpret the data. It should interpret the data for them.</p><p>In my experience, the most valuable reporting systems answer questions before they are asked. They detect when certain routes consistently cause fatigue. They flag when preferred hotels are driving unnecessary taxi costs. They reveal when policy caps are out of sync with market reality. They show which trips next quarter will exceed projected budgets based on current patterns. They identify where exceptions are symptoms of design flaws rather than traveler missteps.</p><p>This shift requires more than data aggregation. It requires <strong>behavioural intelligence</strong>, <strong>contextual interpretation</strong>, and <strong>predictive modeling</strong>. The industry is already moving in this direction. According to FCM’s 2025 benchmark report, <strong>sixty eight percent</strong> of travel leaders now expect reporting tools to provide predictive scenarios instead of static histories. This reflects a broader recognition that data only has value when it <strong>changes decisions</strong>.</p><p>In my opinion, helpful reporting should feel like a conversation. Managers should be able to ask simple questions like <strong>“Where are our highest risk trips this month?”</strong> or <strong>“Which departments are trending over budget?”</strong> and receive clear, actionable responses. They should not need to export data, create pivot tables, or chase different systems for missing information. The system should unify data across travel, expense, risk, sustainability, and finance to create a single narrative.</p><p>The best reporting tools will also reveal <strong>root causes</strong> rather than surface level symptoms. If exceptions are increasing, they will identify why. If costs are rising, they will separate structural changes from behavioural ones. If sustainability metrics are improving, they will link that improvement to specific decisions within the workflow.</p><p>This is how reporting becomes a strategic asset instead of a compliance requirement. It shapes policy intelligently, improves traveler experience, and helps organisations invest in journeys that matter. When done well, reporting not only informs but transforms. It becomes the backbone of continuous improvement across the entire travel ecosystem.</p><h3>My 2 Cents</h3><p>In my opinion, the real opportunity in corporate travel is not about adding more features, more dashboards, or more automation. It is about restoring <strong>judgment at scale</strong>. Technology can take away repetitive tasks. It can connect data that was previously fragmented. It can surface patterns that humans would not have seen on their own. But it cannot understand the deeper <strong>why</strong> behind a journey. It cannot recognise the emotional stakes of a negotiation or the importance of arriving rested for a critical presentation.</p><p>This is why the future of corporate travel is not a contest between humans and machines. It is a partnership. AI brings structure, speed, memory, and predictive insight. Humans bring meaning, empathy, and context. The systems that will define the industry are the ones that blend these strengths rather than treat them as competing forces.</p><p>In my view, progress will not be measured by how much technology we deploy. It will be measured by <strong>how wisely we use it</strong>. A program that automates everything without understanding context is not intelligent. A program that overwhelms users with options is not modern. A program that enforces policy without explaining the reasoning is not trustworthy. Intelligence without humanity feels mechanical. Humanity without intelligence feels inefficient. The future requires both.</p><p>This is why I believe the next wave of travel platforms must be <strong>intent aware</strong>, <strong>explainable</strong>, <strong>contextual</strong>, and <strong>genuinely human centered</strong>. They must begin with the purpose of the trip. They must remove friction through agentic intelligence. They must learn from deviations rather than punish them. They must show cost, carbon, risk, and comfort in one coherent view. And above all, they must treat travel as a strategic capability that supports people rather than an administrative task that constrains them.</p><p>When systems start with intent, everything becomes clearer. Travelers make better choices. Managers trust the process. Finance teams see fewer exceptions. Duty of care teams gain real time visibility. Executives understand the value of travel more easily. In my experience, alignment is not achieved by enforcing rules. It is achieved by building systems that make the right decisions the easiest decisions.</p><p>This is why my final view is simple. Travel should feel <strong>intelligent</strong>, <strong>empathetic</strong>, and <strong>worth the journey</strong>. It should help people arrive ready to contribute, not drained from navigating outdated processes. It should strengthen relationships, not weaken them. When we design for people first, everything else aligns naturally.</p><p>In my opinion, the path forward is not complicated. Begin with intent. Use AI to remove friction. Learn continuously from behaviour. Explain every recommendation clearly. Treat travel as a human experience, not a cost line. When we do that, travel becomes what it should have always been. Not an argument. An advantage.</p><p>I will end the paper is one of the most popular quirks or quotes in boardrooms for a reason and that is <em>“Hope is not a strategy.”</em></p><h3>References</h3><ul><li>Global Business Travel Association. Business Travel Outlook 2025.</li><li>Flight Centre Travel Group. Global Business Travel Report 2024 to 2025.</li><li>FCM Consulting. Quarterly Trend Report 2024.</li><li>American Express Global Business Travel. Corporate Travel and AI Transformation Insights.</li><li>SAP Concur. Global Business Travel Survey 2024.</li><li>Trip.com Group. Corporate and Leisure Travel Recovery Data 2024.</li><li>Booking.com. Travel Technology and Personalization Trends.</li><li>Hopper. Fintech Features and Predictive Pricing Research.</li><li>Capital One Travel. Strategic Expansion into Travel Technology.</li><li>Expedia Group. Traveler Behavior and Platform Intelligence Report.</li><li>PwC, KAYAK, and Blockskye. Reinventing Corporate Travel Through Direct Booking Infrastructure.</li><li>Amex GBT and SAP Concur. Integration and Experience Partnership Report.</li><li>Web In Travel. Asia’s Corporate Travel Recovery and AI Integration.</li><li>NextBigFuture. Elon Musk Describes the AI Node Concept.</li><li>Deloitte. Corporate Travel Trends and Outlook 2025.</li><li>Plaid and Tink. Open Banking Integration and Risk Insights.</li><li>ProQuest and EBSCO. Discovery Layer Design and Information Retrieval Studies.</li><li>PhocusWire. AI Adoption in Corporate Travel 2025.</li><li>Microsoft. Semantic Search and AI Assistant Patterns.</li><li>Anthropic. Agentic AI and Autonomous System Design Principles.</li><li>OpenAI. Intelligent Agent Reasoning and Execution Models.</li><li>McLuhan, Marshall. Understanding Media.</li><li>Drucker, Peter. Selected Writings on Management and Decision Making.</li><li>Einstein, Albert. Selected Reflections on Learning and Insight.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3945418aff97" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Obsolete or Novel in the AI Age]]></title>
            <link>https://medium.com/@prabu.sangar/obsolete-or-novel-in-the-ai-age-39c4926e252a?source=rss-ed44d3335f1d------2</link>
            <guid isPermaLink="false">https://medium.com/p/39c4926e252a</guid>
            <category><![CDATA[education-technology]]></category>
            <category><![CDATA[future-of-work]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[libraries]]></category>
            <category><![CDATA[digital-transformation]]></category>
            <dc:creator><![CDATA[Prabu Sangar]]></dc:creator>
            <pubDate>Sat, 01 Nov 2025 13:04:49 GMT</pubDate>
            <atom:updated>2025-11-01T13:04:49.316Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Vj8LEMkpj7bRSjxi" /><figcaption>Photo by <a href="https://unsplash.com/@guzimak?utm_source=medium&amp;utm_medium=referral">Guzel Maksutova</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h3>Libraries, Age Old Story</h3><p>For centuries, libraries have stood as physical and intellectual anchors in society. They were once the only places where knowledge was stored, organised, and shared. When Johannes Gutenberg introduced the printing press in the fifteenth century, scholars feared that the mass production of books might erode the value of curated collections. Yet librarians adapted, shifting from mere custodianship to roles that emphasised selection, preservation, and education. The same pattern repeated itself with every wave of change. When card catalogues were replaced by digital systems, when online databases began competing with physical stacks, and when search engines put information a click away, the role of the librarian did not disappear. It transformed.</p><p>Today, we are witnessing the next inflection point. Artificial intelligence has entered the library domain, and its presence is impossible to ignore. Large language models can generate summaries of complex texts in seconds. Retrieval-augmented systems can scan millions of academic papers and return tailored results far beyond the capabilities of manual searching. Natural language interfaces allow students to ask open-ended questions and receive structured, relevant answers without needing to understand Boolean logic or controlled vocabularies. The reference desk, once a cornerstone of the library experience, now competes with chatbots that operate continuously and at scale.</p><p>The question that naturally arises is whether librarians remain necessary in this new landscape. The simplest tasks, such as cataloguing, shelving, and basic reference support, are already being automated. As referenced by the American Library Association, routine metadata creation can now be performed by AI tools with remarkable speed and accuracy.</p><p>In Singapore, for example, the National Library Board has deployed autonomous shelf-reading robots that scan entire floors overnight, identifying misplaced books with near-perfect precision. In the United States, universities such as San José State are experimenting with retrieval-augmented chatbots that can answer student queries about research methods, citation styles, and database access without any human intervention.</p><p>Yet these developments do not render Librarians irrelevant. Instead, they redefine what librarianship means. Routine tasks may fade into the background, but the demand for judgment, context, and ethical decision-making has never been greater. Professional Librarians, for instance, are now expected to lead conversations on digital privacy, intellectual freedom, and the ethics of AI in knowledge environments. Teacher Librarians are not simply supporting classroom instruction, they are guiding students through the complex process of evaluating machine-generated content, recognising bias, and cultivating digital literacy.</p><p>The situation mirrors broader workforce trends. In my previous writing on human resources and AI, I argued that automation excels at structured execution but falls short where empathy, governance, and strategic decision-making are required. The same is true in libraries. As Carl Sagan once wrote, “We live in a society exquisitely dependent on science and technology, in which hardly anyone knows anything about science and technology.” Librarians, particularly those with pedagogical and strategic responsibilities are now essential translators of that science and technology for the communities they serve.</p><p>History shows that librarianship is not defined by the tools it uses but by the purpose it serves. In an era where information is abundant but trust is fragile, libraries are evolving from repositories into spaces of interpretation, critical thinking, and human connection. AI is not a threat to that mission. It is a catalyst forcing the profession to rethink, reimagine, and reaffirm its value.</p><h3>Redefining Roles</h3><p>As artificial intelligence transforms the way knowledge is organised, discovered, and delivered, it is tempting to treat “librarian” as a single, uniform category. In reality, the profession is layered, with each role carrying distinct responsibilities, expectations, and vulnerabilities to automation. Understanding how these layers are shifting is essential to understanding the future of the field. I do want to caveat that this is based on my experience working as a Librarian for at least 8 years, where I got an opportunity to work alongside great minds in both academic and school libraries.</p><h3>Library Assistants</h3><p>Library Assistants have historically formed the operational backbone of library services. Their responsibilities often include shelving books, organising materials, registering patrons, managing check-ins and check-outs, preparing thematic displays, and assisting with basic queries. They are the first faces most users encounter, providing the day-to-day support that keeps library spaces running smoothly.</p><p>Yet it is precisely this focus on routine that makes the assistant role most susceptible to automation. Autonomous shelf-reading systems, already deployed by Singapore’s National Library Board, can scan thousands of titles overnight with near-perfect accuracy. AI-powered chatbots now answer basic user questions, process account renewals, and send reminders. Even thematic display suggestions, once the product of human creativity are being supplemented by algorithms that analyse borrowing trends to propose relevant topics.</p><p>But this does not mean the role will disappear. Instead, its focus will shift. The assistant of the future will be less about routine circulation and more about interpersonal engagement. They will act as facilitators, helping patrons navigate digital services, use AI research tools, and access online resources. The task will move from performing processes to supporting people, an evolution that reinforces, rather than diminishes, their value.</p><h3>Library Officers</h3><p>Library Officers sit between operational support and professional leadership. They often manage small teams of assistants, oversee daily operations, coordinate events, and liaise with schools or community groups. They may also curate thematic displays, manage reading programmes, and handle acquisition requests under the guidance of a Professional Librarian.</p><p>In the AI era, many of these responsibilities are being re-defined. Event promotion, for example, is now often driven by predictive analytics that target specific audience segments. Acquisition decisions are informed by usage data and algorithmic recommendations. Yet officers remain crucial, because while machines can identify patterns, they cannot interpret them in a community context. It still takes human judgment to understand why a particular theme resonates in one school but not another, or how a programme might need to be adapted to suit a local audience.</p><p>The officer’s future lies in experience design. Rather than simply executing programmes, they will shape how patrons interact with library services and this will be for both physical and digital resources. They will integrate AI tools into programming, design hybrid experiences that combine online and in-person engagement, and use data to make user journeys more intuitive and inclusive.</p><h3>Professional Librarians</h3><p>At the top of the professional hierarchy are librarians whose responsibilities extend far beyond daily operations. They shape collection development strategies, negotiate digital access, design literacy curricula, support complex research, and set policies for information governance. They are also the public face of the institution in conversations about intellectual freedom, digital ethics, and technology adoption.</p><p>AI is transforming these responsibilities, but not replacing them. Algorithms can recommend acquisitions, but they cannot weigh the cultural, political, or pedagogical implications of those decisions. They can summarise research, but they cannot guide a doctoral candidate through the methodological nuances of a literature review. They can tag content, but they cannot decide what representation, diversity, or equity should mean within a collection.</p><p>Professional Librarians will increasingly act as strategists and ethical stewards. They will guide institutions on how to deploy AI responsibly, how to audit algorithmic outputs for bias, and how to integrate machine intelligence into educational and research workflows. They will also become educators, teaching not just how to find information but how to question it. As referenced by the International Federation of Library Associations, “The librarian’s responsibility is not only to facilitate access to knowledge but to cultivate the capacity for critical evaluation”.</p><h3>Teacher Librarians</h3><p>Teacher Librarians occupy a particularly pivotal space in schools. They have bridged pedagogy and information science, collaborating with teachers to design inquiry-based lessons, teaching research and citation skills, and embedding literacy outcomes across the curriculum. AI expands both the scope and the urgency of this work.</p><p>Students now encounter machine-generated content daily, often without the skills to evaluate its reliability or understand its limitations. Teacher Librarians are uniquely positioned to address this gap. They can teach students how generative models work, how to identify bias in algorithmic outputs, and how to cross-verify information from multiple sources. They can also work alongside teachers to design assignments that use AI tools constructively, not as shortcuts, but as catalysts for deeper inquiry.</p><p>In many ways, the Teacher Librarian’s role is evolving into that of a “learning architect”, designing environments where technology supports curiosity rather than replacing it. As Albert Einstein famously said, “The mind that opens to a new idea never returns to its original size”. Teacher Librarians are guiding students through those new ideas, ensuring that AI becomes a tool for intellectual expansion rather than complacency.</p><h3>Technology and Tools</h3><p>Technology is not replacing librarianship. It is rewriting its boundaries. The systems being deployed today are not simple upgrades; they are changing how decisions are made, how services are delivered, and where human expertise is most needed. To understand the scale of that shift, we must look beyond tools and examine how they alter the balance between automation and human judgment.</p><h3>From Manual Precision to Machine Scale</h3><p>Metadata has always been the invisible scaffolding that makes a library work. Every subject heading, every call number, every keyword is a decision that shapes how knowledge is discovered. In the past, this was slow, meticulous work handled by trained professionals. Now, AI-driven cataloguing systems can suggest subject headings, summarise content, and even propose classification codes in minutes.</p><p>The practical impact is profound. A single professional librarian once spent hours cataloguing a collection of new materials. Today, an AI-assisted workflow can complete the same task in a fraction of the time, with humans focusing on quality control and contextual nuance. As noted by the Library of Congress, machine-assisted metadata creation is no longer experimental; it is becoming part of daily operations.</p><p>This change frees professionals to focus on higher-value tasks, but it also raises strategic questions. Who is responsible if an algorithm embeds bias into a classification system? How do librarians ensure that cultural context and diverse perspectives are preserved when machines prioritise statistical patterns over human meaning? These are not technical issues; they are professional ones. And they demand expertise that machines cannot supply.</p><h3>From Search to Synthesis</h3><p>The library catalogue has evolved from a tool for locating books to a platform for generating insights. Retrieval-augmented generation (RAG) systems are at the forefront of this shift. Rather than presenting a list of search results, they retrieve relevant sources, synthesise them, and present a cohesive summary with citations. This is not theoretical. At San José State University, a pilot RAG service now answers complex questions using curated databases, saving students hours of preliminary research.</p><p>But here too, human oversight remains essential. RAG can summarise, but it cannot judge the reliability of a source or understand the nuance of historical interpretation. It cannot know when a student is relying too heavily on secondary literature or missing a key primary source. Librarians must now become research coaches, teaching users how to question machine outputs and deepen their inquiries. The work is less about finding information and more about building understanding.</p><h3>From Data Crunching to Predictive Collection Development</h3><p>Collection development has always involved a balance of demand, diversity, and mission. Librarians must decide not just what is popular but what is important. Predictive analytics tools are changing that process. They can analyse borrowing trends, citation networks, and even social media conversations to recommend acquisitions. They can flag underused materials for review or predict future interest in emerging fields.</p><p>These capabilities are powerful, but they come with risk. Popularity does not always align with pedagogical value. Predictive models may reinforce existing biases or overlook marginalised perspectives. As one academic librarian at a major US university observed, “Algorithms are good at telling us what people are already reading. They are less good at telling us what they <em>should</em> be reading”. The professional librarian’s role is to interpret the data, challenge its assumptions, and ensure that collections remain intellectually diverse and socially relevant.</p><h3>From Tasks to Strategy</h3><p>Agentic AI, systems capable of performing multi-step tasks without constant human input and this represents the next leap forward. A well-configured agent can manage acquisition workflows, monitor licence renewals, generate usage reports, and even propose new programmes based on community trends. In Singapore, autonomous shelf-reading robots have already transformed how physical collections are maintained, scanning thousands of titles overnight with near-perfect accuracy.</p><p>The shift is not just about efficiency. It is about redefining authority. As routine decisions become automated, librarians must take on new responsibilities: setting ethical boundaries, auditing system behaviour, and designing workflows that align with institutional priorities. As Morpheus tells Neo in <em>The Matrix</em>, “There is a difference between knowing the path and walking the path”. AI can show the path, but librarians must still choose how to walk it.</p><h3>Humanising the Digital Interface</h3><p>Perhaps the most visible transformation is occurring at the patron interface. Virtual assistants now handle account management, reading recommendations, and basic inquiries. Adaptive learning systems personalise reading lists for students. Conversational agents explain citation formats or guide users through research strategies. These tools enhance access and convenience, but they also risk reducing the human connection that defines the library experience.</p><p>This is where librarians are indispensable. They can contextualise recommendations, understand the emotional dimensions of learning, and build trust. They can design programmes that teach not only how to use AI but how to think critically about it. As Nelson Mandela once said, “Education is the most powerful weapon which you can use to change the world”. Librarians are not just deploying technology, they are shaping how people learn to use it responsibly.</p><h3>The Human Imperative</h3><p>Technology can automate tasks, but it cannot replicate judgment, empathy, or institutional memory. It cannot balance competing priorities, defend intellectual freedom, or decide when access must yield to ethics. These responsibilities will define the next generation of librarianship. AI will be the engine, but librarians will remain the conscience.</p><h3>What Will Change and What Will Remain</h3><p>Artificial intelligence will not erase librarianship in the next two years, but it will rearrange its structure in ways that are already becoming visible. Some tasks will disappear quietly into the background. Others will become more important than ever. And some roles will transform so significantly that they will barely resemble what they once were. The choices librarians make now will determine whether they remain at the centre of this change or find themselves sidelined by it.</p><h3>The Tasks That Will Fade</h3><p>The most obvious shift will be the continued decline of repetitive, process-driven work. Routine shelving, basic circulation, and simple reference queries are already being automated in many libraries. Autonomous shelf-reading robots in Singapore and machine-assisted cataloguing in large academic systems show that what was once labour-intensive is now becoming invisible. Even reading recommendations, once a point of personal connection between staff and patrons, are being generated algorithmically based on borrowing histories and demographic patterns.</p><p>This does not mean these tasks are unimportant. It means they will no longer define the human role. They will become infrastructure, handled by systems that operate quietly in the background. Staff who continue to define their value by performing them will find their roles shrinking.</p><h3>The Responsibilities That Will Expand</h3><p>Other responsibilities will grow significantly in importance. The biggest among them is information literacy education. As generative models proliferate, misinformation and bias will become more complex and harder to detect. Librarians, will be on the front line of teaching students and patrons how to critically evaluate AI-generated content, verify sources, and understand the mechanics behind algorithmic recommendations. This will require not only technical literacy but also pedagogical skill.</p><p>Policy development and governance will also become central. Institutions will need professionals who can design ethical frameworks for AI use, manage privacy concerns, and ensure that algorithms align with values such as equity and intellectual freedom. Librarians will increasingly be part of interdisciplinary teams working alongside technologists, administrators, and educators to shape how AI is implemented and monitored.</p><p>Collection development will remain vital but will shift in character. Predictive analytics will support decision-making, but human oversight will remain necessary to ensure that collections reflect diversity, address equity gaps, and serve pedagogical needs rather than merely following usage patterns.</p><h3>Roles That Will Evolve</h3><p><strong>Library Assistants</strong> will transition from operational support to service facilitation. Instead of shelving and checkouts, they will help patrons navigate digital platforms, use AI-powered research tools, and access remote resources. Their interpersonal skills will become more valuable than their technical ones.</p><p><strong>Library Officers</strong> will evolve into experience designers. They will orchestrate how physical and digital services interact, integrating technology into programming and community engagement. They will use analytics to shape user journeys, but they will also ensure that these journeys remain inclusive and human-centred.</p><p><strong>Professional Librarians</strong> will move deeper into strategy, policy, and institutional leadership. They will be responsible not just for curating content but for shaping how knowledge is produced, interpreted, and used. Their work will be increasingly interdisciplinary, drawing on fields such as data ethics, educational psychology, and digital humanities.</p><p><strong>Teacher Librarians</strong> will become architects of learning ecosystems. Their classrooms will extend beyond physical walls, encompassing AI-driven adaptive learning platforms and collaborative research spaces. They will teach not only information literacy but also <em>algorithmic literacy</em> — the ability to understand, question, and shape how technology influences knowledge.</p><h3>Redefining the Future of Librarianship</h3><p>The next two years are not about survival. They are about reinvention. Libraries must move beyond the question of whether technology will replace them and instead ask how human judgment, institutional purpose, and ethical stewardship can shape the future. The rise of artificial intelligence is not dismantling librarianship. It is creating new spaces for relevance, but only for those who are prepared to evolve.</p><h3>Rethinking Workforce Models</h3><p>The workforce that powers libraries will look fundamentally different in the AI era. Traditional staffing models, built around fixed roles and predictable tasks, will give way to outcome-driven approaches that prioritise expertise, agility, and measurable impact.</p><p>Business Process Outsourcing (BPO) will continue to support operations, but its nature is changing. In Singapore, schools would be shifting from hiring Library Assistants and Library Officers for fixed schedules to contracting external teams for specific deliverables and this would entail curriculum-linked thematic displays, parent literacy sessions, or targeted engagement campaigns. In this “bounty model,” support is deployed when needed, reducing cost while improving flexibility.</p><p>In the United States, public and academic libraries have extended outsourcing beyond logistics to specialised services. External partners now assist with metadata enhancement, corpus curation for retrieval-augmented generation systems, and analytics-driven insights into user behaviour. These arrangements allow internal teams to focus on higher-order priorities: pedagogy, governance, and strategic programme design.</p><p>The human workforce will remain central, but its focus will shift. Assistants and Officers will evolve into facilitators, guiding patrons through increasingly complex digital systems and helping them apply AI tools effectively. Professional Librarians and Teacher Librarians will become architects of strategy, designing frameworks for ethical technology use, aligning library services with institutional goals, and shaping long-term knowledge ecosystems. The role of humans will no longer be to operate the machine, but to ensure it operates in service of learning, equity, and trust.</p><h3>New Professional Identities</h3><p>As library work diversifies, entirely new professional identities are emerging. These roles reflect not just new tasks, but new responsibilities that sit at the intersection of information, ethics, and technology.</p><p>One of the most significant is the <strong>AI Literacy Specialist</strong>. This role focuses on teaching both staff and users how AI works, how to interpret its outputs, and how to apply it critically in research and learning. Their work is not limited to workshops, it extends into curriculum design, where they collaborate with teachers and faculty to embed algorithmic literacy into everyday instruction.</p><p>Another pivotal role is the <strong>Algorithm Governance Librarian</strong>, tasked with auditing the systems that underpin discovery platforms, metadata generation, and content recommendations. They ensure compliance with privacy laws, evaluate algorithmic bias, and document decision-making logic. These librarians are the safeguards of intellectual integrity in a machine-augmented landscape.</p><p>The <strong>Digital Engagement Librarian</strong> will focus on designing hybrid experiences that combine the scale of technology with the nuance of human facilitation. They will create programmes that integrate chatbots, adaptive learning tools, and retrieval-augmented research assistants while maintaining a human-centred approach.</p><p>These roles are already taking shape. Stanford University’s AI &amp; Data Literacy initiative has created librarian-led modules on algorithmic bias, while Singapore’s National Library Board is expanding its digital engagement capabilities to include AI-driven education programmes. The shift is clear: librarians are moving from facilitators of access to curators of context and custodians of trust.</p><h3>Competencies That Will Matter Most</h3><p>The librarians of tomorrow will need deeper and broader skills than their predecessors. These competencies are not optional, they will define who remains relevant in an AI-driven world.</p><h4>Data and Analytical Literacy</h4><p>Librarians must learn to interpret usage analytics, evaluate recommendation patterns, and explain how algorithms make decisions. This knowledge will allow them to design better services and teach patrons to navigate digital complexity.</p><h4>Algorithmic Understanding</h4><p>Beyond using tools, librarians must understand how they work — and where they fail. This includes recognising hallucinations, identifying bias, and explaining the limitations of machine reasoning.</p><h4><strong>Policy and Governance Expertise</strong></h4><p>Knowledge of data protection laws, copyright frameworks, and AI regulatory standards will become essential. Librarians must be capable of drafting policies that balance innovation with ethical responsibility.</p><h4><strong>Pedagogical Design Skills</strong></h4><p>Especially for Teacher Librarians, the future demands the ability to integrate AI into curriculum meaningfully. This includes designing assignments that encourage students to challenge machine outputs, not just consume them.</p><h4><strong>Change Leadership</strong></h4><p>Finally, librarians must develop the capacity to lead institutional transformation, advocating for responsible AI use, training peers, and guiding their organisations through periods of technological transition.</p><p>As Peter Drucker observed, “The greatest danger in times of turbulence is not the turbulence; it is to act with yesterday’s logic”. Building these competencies is how libraries avoid that trap.</p><h3>The Vendor Imperative</h3><p>The future of librarianship is not shaped by librarians alone. It is equally shaped by the systems that power their work. Library system vendors and platform developers must evolve from service providers into strategic partners that enable, rather than constrain, institutional ambition.</p><p>Closed ecosystems that limit integration will become liabilities. Vendors must design platforms with open APIs, modular architectures, and compatibility with retrieval-augmented generation tools. Libraries must be able to plug in new capabilities without rebuilding their core infrastructure.</p><p>As AI begins to influence acquisitions, recommendations, and discovery results, explainability is non-negotiable. Vendors must offer audit trails and documentation that allow librarians to understand and challenge machine decisions.</p><p>Technology should amplify human judgment, not bypass it. Systems must allow librarians to review, approve, or override AI outputs before they are shared with patrons.</p><p>Privacy compliance, consent management, and algorithmic bias detection must become standard features. These capabilities are essential for maintaining public trust and meeting evolving legal requirements.</p><p>Vendors must also invest in training librarians, not just on how to use their systems, but on AI literacy, risk assessment, and ethical deployment. The most successful partnerships will be those that build human capacity alongside technological capacity.</p><p>As noted by the Library of Congress, “The future of library technology lies not in replacing librarians but in amplifying their capacity to act”. Vendors that embrace this principle will be essential to the next generation of library innovation.</p><h3>A Practical Path Forward</h3><p>Transforming libraries for the AI era is not a single project. It is a continuous process of experimentation, reflection, and recalibration. The most successful institutions will approach this journey in deliberate stages: first, by defining clear strategic priorities; then, by building internal capacity; and finally, by iterating through small, evidence-based pilots before scaling.</p><p>Partnerships will be central. Collaboration with vendors, universities, policymakers, and private-sector experts will accelerate capability building and reduce risk. Shared standards, open platforms, and co-developed tools will ensure libraries remain agile as technologies evolve.</p><p>Most importantly, the human element must remain at the centre of this transformation. As <em>The Matrix</em> reminds us, “There’s a difference between knowing the path and walking the path”. Librarians now know the path. The question is whether they will walk it, not as caretakers of tradition, but as architects of a future where knowledge is more accessible, more equitable, and more human than ever before.</p><h3>References</h3><ol><li>Library of Congress. (2024). <em>Metadata and Machine Learning in Libraries: Emerging Standards and Practices</em>. Washington, DC: Library of Congress Publications.</li><li>American Library Association. (2023). <em>Artificial Intelligence and the Future of Libraries: Opportunities, Risks, and Ethical Considerations</em>. Chicago: ALA Press.</li><li>International Federation of Library Associations and Institutions (IFLA). (2023). <em>Guidelines for AI in Library and Information Services</em>. The Hague: IFLA.</li><li>National Library Board Singapore. (2024). <em>Annual Technology Report: Automation and Intelligent Systems in Public Libraries</em>. Singapore: NLB Publications.</li><li>San José State University Library. (2024). <em>AI and Academic Research: Case Study on Retrieval-Augmented Generation in Higher Education Libraries</em>. San José, CA: SJSU Research Series.</li><li>Brynjolfsson, E., et al. (2023). <em>AI Adoption in Knowledge Work: Field Studies and Implications for Learning Institutions</em>. MIT Initiative on the Digital Economy.</li><li>Mandela, N. (1994). <em>Long Walk to Freedom</em>. Johannesburg: Macdonald Purnell.</li><li>Einstein, A. (1946). <em>Address at Princeton on Education and Change</em>. Princeton University Archives.</li><li>Hawking, S. (1996). <em>Black Holes and Baby Universes and Other Essays</em>. New York: Bantam Books.</li><li>Freire, P. (1970). <em>Pedagogy of the Oppressed</em>. New York: Continuum.</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=39c4926e252a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI, A Tool to Our Teammate]]></title>
            <link>https://medium.com/@prabu.sangar/ai-a-tool-to-our-teammate-54e7fc653e0e?source=rss-ed44d3335f1d------2</link>
            <guid isPermaLink="false">https://medium.com/p/54e7fc653e0e</guid>
            <category><![CDATA[travel-technology]]></category>
            <category><![CDATA[digital-transformation]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[agentic-ai]]></category>
            <category><![CDATA[future-of-work]]></category>
            <dc:creator><![CDATA[Prabu Sangar]]></dc:creator>
            <pubDate>Tue, 07 Oct 2025 20:58:52 GMT</pubDate>
            <atom:updated>2025-10-07T20:58:52.880Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*UDQ8qytBCYwRUznI" /><figcaption>Photo by <a href="https://unsplash.com/@vladhilitanu?utm_source=medium&amp;utm_medium=referral">Vlad Hilitanu</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>There is a moment in every technological wave when novelty becomes necessity. It doesn’t usually arrive with fireworks. Instead, it sneaks in quietly, until one day, we realize we’ve crossed a line.<br>With GPT-5, we are living through that line-crossing moment.</p><p>Previous generations of AI were reactive. They wrote code, drafted emails, completed sentences. But they waited for instruction. GPT-5 is different. It understands objectives, plans steps, executes actions, and learns from outcomes. It does not simply assist. It acts.</p><p>This paper explores the rise of agentic AI systems that operate with purpose. We trace how we got here, why most companies are not ready, and what organizational change is required. Through the lens of the travel industry and with a candid look at leadership reality, we’ll explore how delegation to AI needs more than models. It needs a mindset.</p><h3>LLM Leap and Quantum Behind It</h3><p>To understand agentic AI, we must understand the journey that made it possible.</p><p><strong>GPT-3</strong> arrived in 2020 like a linguistic magician. It could write convincingly and mimic styles with ease. But it was, in truth, an autocomplete machine. No memory. No tools. No understanding of goals. Just language prediction. It struggled with reasoning, hallucinated facts, and could not take actions or use tools. It was a brilliant mimic, but not a doer.</p><p><strong>GPT-4</strong> brought structure to that mimicry. With improved reasoning, greater accuracy, and plugin tools, it could summarize long documents, write basic code, and even pass bar exams. GPT-4 introduced the beginnings of multi-step problem solving and contextual awareness. Yet, it still required hand-holding. It could not initiate, evaluate its own work, or carry memory across sessions in any functional way.</p><p><strong>GPT-5</strong>, launched in 2025, marks a new epoch. It brings goal-oriented behavior, memory persistence, recursive planning, and autonomy. The model can now take a high-level instruction and deconstruct it into a sequence of actionable tasks. It can use APIs, tools, and data sources without direct human prompting. And it can learn preferences over time.</p><p>As an example, GPT-5 can:</p><ul><li>Plan a multi-city business trip across time zones</li><li>Check weather and travel disruptions</li><li>Follow corporate travel policy on preferred suppliers and classes</li><li>Execute bookings, send confirmation emails, and prepare post-trip reports</li></ul><blockquote>All without micromanagement.</blockquote><p>According to Accenture’s 2025 AI Pulse Report, 71 percent of Fortune 1000 enterprises that piloted GPT-5-based tools said they expect to integrate agentic AI into critical workflows within the next 24 months.</p><p>A use case from one Fortune 500 logistics company showed a 47 percent reduction in operational task load after implementing a GPT-5-based agent to manage procurement and shipment routing queries across three continents. The system flagged pricing anomalies, cross-validated customs declarations, and routed exceptions to human leads, all without formal prompts.</p><p>To drive the point further, in a report by McKinsey published in June 2025, businesses that had moved to agentic AI for back-office operations reported a 3.8 times higher ROI on automation spend than those still using traditional RPA.</p><p>In addition, a case study by the World Economic Forum detailed how a leading European energy firm used GPT-5 to automate internal policy documentation, stakeholder communications, and compliance monitoring. The result was a 58 percent time reduction in legal reporting tasks and improved interdepartmental transparency, as cross-functional teams could query and understand compliance frameworks in natural language through an agent.</p><p>What is most compelling is not just the savings or productivity boosts. It is the capacity for AI to collaborate, learn and remember, which transforms it from a utility to a workforce partner. This evolution is not cosmetic. It is structural.</p><h3>What Is Agentic AI?</h3><p>Agentic AI refers to systems that go beyond passive interaction to proactively plan and carry out tasks in pursuit of goals. These systems are not just predictive, but purposive. They behave like autonomous collaborators with embedded intent and the ability to adapt in real time. They are capable of engaging in recursive reasoning, making decisions without human micro-intervention, and updating their knowledge base on the fly.</p><p>To put it plainly, if GPT-3 was a writing assistant, GPT-5 is a project manager.</p><p>Agentic AI is grounded in the theory of cognitive architectures, systems that replicate aspects of human-like decision-making, including perception, memory, planning, and execution. Leading models can now be configured to interface directly with third-party systems like ERPs, CRMs, or even call center software. They can track performance, adjust output formats, and even apply learned business rules across tasks. These are not isolated features, they form an ecosystem of intelligent operations.</p><p>A recent Deloitte AI Index showed that companies implementing agentic capabilities saw 5 to 7 percent revenue uplift in customer-facing functions, with the highest returns in telecom, logistics, and travel. The reason is that these systems shorten the feedback loop between a problem and its resolution. Faster, more accurate action leads to improved SLAs, customer satisfaction, and lower costs.</p><p>One example is an Australian telco deployed an agentic AI to handle SIM card activation, identity verification, and port-in authorization. In 90 days, customer onboarding time dropped from 36 hours to under 20 minutes. Error rates fell by 82 percent. The AI adapted its flow based on user behavior, automatically escalating outliers to fraud teams without manual logic scripting.</p><p>In the same way, enterprise travel is rife with repetitive decision trees, fare types, seat availability, hotel policy restrictions. These decisions, once mapped and trained, are well within the wheelhouse of a GPT-5-enabled agent.</p><p>This is no longer about replacing tasks. It is about re-architecting how we think about operational roles.</p><h3>Why Travel Is the Perfect Testbed?</h3><p>The travel industry is an ecosystem of interdependent systems, GDSs running on legacy technology techstack, mid-office tools patched together with brittle APIs, and agents expected to reconcile inconsistent fare rules, corporate policy nuances, and loyalty point requests across time zones. There’s perhaps no better environment to test the limits of automation, resilience, and human intervention.</p><p>Unlike finance or healthcare, where regulation shapes digital maturity, travel has remained a patchwork of manual overrides and frontline heroics. Exception handling is not the exception, it is the daily job. Travel consultants manually reprice itineraries, add visa remarks, troubleshoot queue fallouts, or nudge suppliers on bookings. Every task involves reading between the lines.</p><p>This is why GPT-5’s agentic capabilities are especially suited for this sector. In 2025, a pilot program between a global TMC and a travel AI platform demonstrated how agentic AI could absorb and execute over 80 percent of repetitive email-based service requests. These included ticket exchanges, seat preference updates, fare rule lookups, and even PNR shell creation. The impact was transformative. SLA adherence increased by 35 percent across pilot markets. Escalation volumes decreased. And most tellingly, agent stress levels, as measured through post-shift sentiment tracking, improved significantly.</p><p>Here, agentic AI was not replacing the agent, it was augmenting them. When agents are freed from low-value, repetitive work, they can focus on complex cases, nuanced traveller needs, and proactive client engagement. The AI served as a digital apprentice, one that never tires, forgets, or misfiles a remark.</p><p>Moreover, the AI proved to be globally scalable. It could handle a PNR in Mandarin, cross-reference a frequent flyer upgrade rule in Amadeus, then switch context to another content provider query. That multilingual, multi-system fluency is not a “nice to have.” It is essential to survive in a post-pandemic travel environment where clients demand 24/7 service, localization, and personalization at scale.</p><p>And perhaps most notably, the pilot revealed something more profound, and this means travelers didn’t notice. Or rather, they didn’t mind. As long as they received timely, accurate service, the human vs machine debate was irrelevant. In fact, Net Promoter Scores in pilot markets increased by 11 points.</p><p>Agentic AI doesn’t disrupt travel. It relieves it. As one of the Travel Management Companies stated, “Our goal is not to automate for the sake of it, but to preserve the quality and humanity of travel management, at scale.” This is the future of the industry and it’s talent supported, not sidelined, by technology.</p><p>In a space constrained by outdated back-end infrastructure but pressured by rising traveler expectations, the only way forward is not brute force staffing, it’s intelligent orchestration. Agentic AI, if deployed correctly, offers precisely that.</p><blockquote>“The goal is not to automate for the sake of it, but to preserve the quality and humanity of travel management, at scale.”</blockquote><h3>The Framework for Responsible AI Adoption</h3><p>Yes, I’ve named this framework after myself and admittedly, shamelessly. But only because the letters happened to fit perfectly. And let’s face it, if AI can name its models after Greek philosophers and abstract numbers, why can’t we give a human name to something that actually involves real people?</p><p><strong>PRABU</strong> stands for:</p><ul><li><strong>P</strong>rioritize Data Foundations</li><li><strong>R</strong>edesign Roles, Not Just Workflows</li><li><strong>A</strong>lign on Ethics, Escalation, and Boundaries</li><li><strong>B</strong>uild with Tooling and Training in Tandem</li><li><strong>U</strong>nblock the Feedback Loop</li></ul><p>Let’s break it down.</p><p><strong>Prioritize Data Foundations<br></strong>You cannot deploy agentic AI without first understanding what your systems look like under the hood. Where is your data stored? Is it structured? Can an AI access it reliably? Most enterprises have fragmented architecture which includes travel data in one system, finance in another, and policies in SharePoint folders last updated in 2018. AI needs clean, connected, and context-rich data. This isn’t a technical barrier, it’s a leadership decision.</p><p><strong>Redesign Roles, Not Just Workflows<br></strong>This is where most firms get it wrong. They think deploying AI is about removing a person from a task. It’s not. It’s about redesigning roles around judgment, escalation, exception handling, and collaboration. For instance, when a travel consultant no longer has to manually validate ticketing remarks, what do they do instead? The best organizations reframe job scopes, turning coordinators into curators of insight and escalation managers into customer advocates.</p><p><strong>Align on Ethics, Escalation, and Boundaries<br></strong>Not every decision should be delegated to AI. Companies need a playbook. What happens when an agent recommends a vendor that violates ESG policy? Who checks for bias in credit approval decisions made by an AI in finance? Define guardrails early. Escalation thresholds. Human-in-the-loop workflows. And above all, be transparent. Ethical alignment isn’t a compliance box, it’s a trust engine.</p><p><strong>Build with Tooling and Training in Tandem<br></strong>You can’t just drop GPT-5 into an Excel-heavy department and expect magic. People need to know how to work with it. That means upskilling, not just your developers, but your domain experts, support teams, and yes, your middle managers. Parallel to tool deployment should be training programs that demystify AI, define its limits, and empower teams to collaborate with it, not against it.</p><p><strong>Unblock the Feedback Loop<br></strong>Agentic systems learn. But only if feedback flows. That means you need post-action audits, performance metrics, escalation reviews, and dynamic retraining protocols. Don’t treat your AI like a static tool, treat it like a new team member. One that requires coaching, review, and regular calibration.</p><p>This framework should not be a silver bullet. It’s a provocation. A reminder that successful AI adoption is not a technical upgrade, it’s an organizational rewrite. It demands that leaders take AI seriously enough to go beyond pilot programs and PowerPoints. To build systems of intelligence, not just moments of automation.</p><p>As my colleague Daniel Senyard, SVP of AI Commercialization and Innovation at Flight Centre Travel Group, often reminds us, AI adoption should never be force-fed. Daniel and his team have been at the forefront of translating innovation into relevance, ensuring change feels organic, not imposed.</p><p>True adoption happens when people from their own domains become champions, SMEs who inspire through demonstration, not direction. It is this peer-driven confidence that sustains momentum. When a travel consultant, a finance lead, or a developer adopts AI and finds genuine value, others follow naturally.</p><blockquote>That is the culture of inspired diffusion, not forced compliance.</blockquote><h3>Where Does Travel Go from Here?</h3><p>The travel industry stands on the edge of an AI-induced renaissance, but not without its battles. If agentic AI is the engine, the current travel stack is the rusted chassis still trying to hold it.</p><p>GDSs, for all their scale, remain clunky. They’re grounded in 1970s-era architectures with command-line interfaces that resemble DOS terminals more than modern software. Integration across mid- and back-office tools is often stitched together with manual processes, scripting hacks, or brittle queue logic. The operational inefficiency bleeds into every part of the travel experience.</p><p>This legacy baggage is why agentic AI in travel isn’t just useful, it’s necessary. Imagine an AI that not only recognizes a hotel policy exception coded incorrectly in a PNR, but also backtracks to the client profile, identifies the corporate contract clause, and amends the policy logic across future bookings. That’s not just automation. That’s self-healing infrastructure.</p><p>Consider also the agency side. In an ideal model, AI agents triage requests; flight changes, refunds, queue monitoring, MIS validation, and shell PNR creation while human consultants take on strategic roles. They can build rapport, manage traveler anxieties, offer personalized guidance, or resolve cross-system friction. Think of a digital concierge enabling a human concierge, not replacing them.</p><blockquote>It’s all about the high value tasks, while low value tasks are automated where necessary.</blockquote><p>Yet, amid this transformation, there remains the silent hesitation, <em>the elephant in the room is will AI replace me? </em>All of us think about and I admit I do when I see the use cases and possibilities.</p><p>As <strong>Muhammad Ali</strong> once said, <em>“Don’t count the days. Make the days count”.</em> If we spend our energy fearing displacement, we stop evolving. Instead, rewrite your role. Define how you contribute to your organization’s growth in an AI-driven era. One small step of adaptation is better than standing still.</p><p><strong>In Avenger Ultron</strong> once warned, <em>“When the dust settles, the only thing living in this world will be metal”.</em> The line may be cinematic, but its message is timeless , stagnation is the real threat, not AI.</p><p>My favourite line would be what Daniel had mentioned in our interaction where he shared, the future belongs to those who adapt, not those who resist. Today’s job description is not tomorrow’s. IQ may be finite, but our capacity to learn, grow, and redefine value is infinite.</p><p>A large Asia-Pacific travel management company recently deployed an agentic AI system that handled over 5,000 after-hours booking changes during a typhoon season. Not only did it reroute travelers and notify hotels, it also flagged potential SLA breaches, updated management dashboards in real time, and initiated refund claims without human prompting. NPS scores went up. Consultant overtime went down.</p><p>So what comes next?</p><ol><li>Interoperability upgrades where TMCs must demand APIs from GDSs. Until Amadeus, Sabre, and Travelport offer cleaner, modern interfaces, AI will be duct-taped to mainframes.</li><li>Rebuild processes around AI agents. Not as afterthoughts, but as collaborators.</li><li>Clients must be brought into the loop, explaining what AI is doing on their behalf, and how it supports policy, savings, and traveler well-being.</li></ol><p>For too long, travel technology has been about containment, controlling costs, enforcing policies, minimizing risk. Agentic AI flips the equation. It asks: how do we elevate the travel experience while empowering those who deliver it?</p><p>In a recent conference I attended, I heard this and it stayed with me.</p><blockquote>Travel is not a commodity. It’s a service. A highly human one. AI doesn’t remove that humanity, it protects it, by absorbing the friction. That’s where travel goes from here.</blockquote><h3>Conclusion</h3><p>AI is not coming. It is here. But adoption is not the same as absorption. Just because an organization launches a pilot, deploys a chatbot, or issues a press release about a GPT integration doesn’t mean it has transformed. True transformation is not performative. It is infrastructural.</p><p>So what should companies do now?</p><p><strong>Start by treating AI like a team member, not a tool.</strong> You don’t hire someone without onboarding them. The same must be true for AI. This includes defining roles, responsibilities, escalation paths, KPIs, and audit mechanisms. If your AI is making decisions, who supervises it? If it’s accessing data, who approves access scopes? Treat AI implementation like workforce design.</p><p><strong>Create an AI readiness index for your enterprise.</strong> Inventory your workflows. Identify which tasks are ripe for agentic augmentation, especially those that are repetitive, rules-based, and suffer from error fatigue. Then overlay your findings with system access, data integrity, and escalation mapping. The goal is to not deploy where it’s easiest, but where it matters most.</p><p><strong>Invest in literacy before capability.</strong> A 2025 study by MIT Sloan showed that companies with the highest ROI from AI spent 3.2 times more on training and enablement than their peers. Give your employees context. Help them understand what AI can do and what it cannot. Build trust through clarity, not coercion.</p><p><strong>Formalize your ethical playbook.</strong> Don’t wait for a governance crisis. Build it now. Define when AI should act, when it should wait, and when it should hand off to a human. Set redlines to ensure no AI-generated financial statements without review. No client-facing communication without traceability. And above all, build systems for accountability.</p><p><strong>Get comfortable with transparency.</strong> If your leadership team cannot answer where, how, and why AI is being used across your organization, you are not ready. In a survey published by Fortune in July 2025, only 31 percent of executives could name the core use cases of AI within their own enterprise portfolios. That is unacceptable. AI should not be a black box. It should be a transparent, auditable system.</p><p>And yes, <strong>stop piloting forever.</strong> The pilot phase is over. It’s time to institutionalize. If your organization is still stuck in proof-of-concept purgatory, it is losing ground. The winners are not those with the best models, they’re the ones who turned models into motion.</p><p>The real risk is not AI. It is inertia.</p><p>This is a once-in-a-generation window to redesign how we work. Not to do more with less, but to do better with intention. Agentic AI is not a magic trick. It’s a new way of thinking about responsibility, capability, and scale.</p><blockquote>“AI is the defining technology of our time. But it is not about man versus machine. It is about man with machine.”</blockquote><p>So start. Design. Retrain. Rewire. Reimagine.</p><p>Because the organizations that act now will not only survive this next wave of AI. They will shape it.</p><p>I remain deeply grateful to <strong>Daniel Senyard</strong> (<a href="https://medium.com/u/639e74d4f916">Daniel Senyard</a>) for his insight and guidance in shaping this piece. His clarity that innovation must always remain human-centered is a reminder that progress without empathy is not evolution, it’s erosion. Leaders like Daniel show us that the real power of AI lies not in automation, but in amplification of our intent, intelligence, and integrity.</p><h3>References</h3><ol><li>Harvard Business Review &amp; BCG. (2025). <em>AI Readiness in the Enterprise: A Global Survey</em>. Retrieved from <a href="https://hbr.org/">https://hbr.org</a></li><li>Accenture. (2025). <em>AI Pulse Report: The Future of Enterprise Automation</em>. Retrieved from <a href="https://accenture.com/">https://accenture.com</a></li><li>McKinsey &amp; Company. (2025). <em>The Economic Potential of Generative AI: The Next Productivity Frontier</em>. Retrieved from <a href="https://mckinsey.com/">https://mckinsey.com</a></li><li>Deloitte. (2025). <em>AI Index: Agentic Systems in Global Operations</em>. Retrieved from <a href="https://deloitte.com/">https://deloitte.com</a></li><li>PwC. (2025). <em>Workforce Trust Survey: Navigating AI in the Workplace</em>. Retrieved from <a href="https://pwc.com/">https://pwc.com</a></li><li>Gartner. (2025). <em>CIO Pulse: AI Implementation Trends</em>. Retrieved from <a href="https://gartner.com/">https://gartner.com</a></li><li>World Economic Forum. (2025). <em>Case Study: AI for Compliance in the Energy Sector</em>. Retrieved from <a href="https://weforum.org/">https://weforum.org</a></li><li>MIT Sloan Management Review. (2025). <em>AI Literacy as a Driver of Adoption ROI</em>. Retrieved from <a href="https://sloanreview.mit.edu/">https://sloanreview.mit.edu</a></li><li>Satya Nadella. (2023). Keynote at Microsoft Ignite. Quoted in Forbes Technology Council. Retrieved from <a href="https://forbes.com/">https://forbes.com</a></li><li>Fortune. (2025). <em>AI Transparency and C-Suite Awareness</em>. Retrieved from <a href="https://fortune.com/">https://fortune.com</a></li><li>Forbes Insights. (2024). <em>Generative AI in Customer Service: Real Use Cases</em>. Retrieved from <a href="https://forbes.com/">https://forbes.com</a></li><li>BCD Travel &amp; Oversee AI. (2025). <em>Strategic Partnership Press Release</em>. Retrieved from <a href="https://bcdtravel.com/">https://bcdtravel.com</a></li><li>HBR. (2024). <em>Don’t Automate Just for Efficiency. Automate Intelligently.</em> Retrieved from <a href="https://hbr.org/">https://hbr.org</a></li><li>World Travel &amp; Tourism Council. (2024). <em>AI and the Future of Travel Jobs</em>. Retrieved from <a href="https://wttc.org/">https://wttc.org</a></li><li>Brookings Institution. (2023). <em>Responsible AI Governance: From Concept to Practice</em>. Retrieved from <a href="https://brookings.edu/">https://brookings.edu</a></li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=54e7fc653e0e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Future of HR in the Age of AI]]></title>
            <link>https://medium.com/@prabu.sangar/the-future-of-hr-in-the-age-of-ai-7b1808b94ea5?source=rss-ed44d3335f1d------2</link>
            <guid isPermaLink="false">https://medium.com/p/7b1808b94ea5</guid>
            <category><![CDATA[ai-in-hr-tech]]></category>
            <category><![CDATA[the-future-of-work]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[hr-automation]]></category>
            <dc:creator><![CDATA[Prabu Sangar]]></dc:creator>
            <pubDate>Thu, 02 Oct 2025 06:56:02 GMT</pubDate>
            <atom:updated>2025-10-02T06:56:02.554Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*zfQBlGpCAr2EXiEu" /><figcaption>Photo by <a href="https://unsplash.com/@cytonn_photography?utm_source=medium&amp;utm_medium=referral">Cytonn Photography</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h3>A Note of Thanks</h3><p>Before diving into this piece, I want to extend my gratitude to <strong>Jaslin</strong> for her time and thoughtful contributions. Many of the insights shared here, particularly around predictive analytics, leadership collaboration, and the role of HR in creating psychological safety come directly from our conversation. Without her perspective, this article would not have been possible.</p><h3>The Quiet Fear in HR</h3><p>It began with a simple comment from a consultant: <em>“AI answers my questions faster, but it does not make me feel heard.”</em> That remark captures the unease many employees feel today. HR was once the human heartbeat of organisations, guiding people through their most difficult and defining moments.</p><p>But with AI now writing performance reviews, screening candidates, and advising managers on layoffs, a quiet fear is growing, the million dollar question is that will HR itself become irrelevant?</p><p>According to Gartner, adoption of generative AI in HR has surged from 19 percent in mid-2023 to 61 percent by early 2025. Companies across industries are already hiring HR technologists and AI product leads. This is not a future scenario. It is the present reality. And the question is no longer whether AI can perform HR tasks. The question is whether HR can redefine itself before AI defines it.</p><h3>HR Ops Meets Algorithms</h3><p>HR operations are being reshaped before our eyes. PwC automated candidate screening with AI-powered bots and reported a 30 percent reduction in hiring time. Unilever uses AI-driven video interviews and game-based assessments, slashing recruitment cycles by 75 percent. IBM’s “AskHR” chatbot manages thousands of queries daily, while Accenture relies on AI to generate real-time performance feedback.</p><p>I saw this play out in Singapore when my friend’s company deployed an AI HR assistant to handle staff queries such as overtime rules, travel allowances, and meal reimbursements. The system reduced inbox overload, but one consultant admitted she missed the reassurance of a human confirming, <em>“Yes, you are entitled to that allowance.”</em> Efficiency mattered less than being heard.</p><p>HireBee reports that 60 percent of organisations now manage recruitment end-to-end with AI, while PeopleSpheres notes that 93 percent of employers using HR automation report major time savings. Yet, the human dimension that includes trust, empathy, and context remains beyond automation.</p><p>More recently, AI has crept into everyday HR tasks and this entails drafting letters, preparing handbook updates, suggesting engagement ideas, even summarising staff surveys. These efficiencies free HR from routine paperwork. But responsibility remains human with HR validating compliance with law, checks for tone and fairness, and makes sure policies reflect culture. AI assists. HR decides.</p><h3>When AI Speaks Like a Leader</h3><p>The more unsettling shift is AI stepping into advisory roles. Axios found that 94 percent of managers already rely on AI to assist with promotion and layoff decisions. At one Big 4 office in London, AI dashboards recommended trimming underperforming sectors. The model’s logic was clear. But an HR partner admitted, <em>“AI gave me the decision, but it was HR that had to hold the conversation, look people in the eye, and take accountability.”</em></p><p>It is one thing for AI to process payroll. It is another to have your performance review scripted by a machine. Would you feel supported, or replaced?</p><p>One truth has always stayed with me, <em>“Anybody can offer advice, but it takes a special someone in HR, who knows us and understands us, to provide guidance that truly resonates.”</em> That difference is that the human ability to connect with empathy and context is where AI cannot compete.</p><p>In Singapore, an AI tool flagged consultants logging excessive overtime and suggested reducing headcount in other regions to balance costs. On paper, the numbers worked. Yet HR leaders paused, recognising cultural nuance given in some markets, staff quietly accept long hours without complaint. Only a human HR manager could interpret what the data could not.</p><p>This is why psychological safety matters. When AI makes a recommendation, employees need reassurance that HR will review it, explain it, and make the final decision with compassion.</p><h3>From Cost-Cutter to Culture Builder</h3><p>AI has changed HR budgets too. IBM famously cut 40 percent of its HR spend through automation. But Deloitte’s Human Capital Trends survey shows the best-performing organisations are those that blend AI efficiency with human-centered governance.</p><p>I saw two different responses. In Singapore, a regional HR head rejected an AI scheduling tool’s suggestion to relocate staff from Jakarta to Sydney. <em>“AI does not know what it means to uproot a family,”</em> she said. <em>“It only knows margins.”</em></p><p>At Deloitte in New York, predictive AI flagged surplus auditors and recommended reducing staff. Instead, HR leaders reskilled them into ESG assurance, a field in demand. Efficiency was achieved, but so was trust.</p><blockquote>The message is simple. AI delivers numbers. HR delivers meaning.</blockquote><h3>The Next 1–3 Years: Automating the Routine</h3><p>AI will increasingly handle transactional tasks, screening resumes, answering routine policy questions, scheduling. In Singapore, chatbots now reply instantly to questions like, <em>“Am I entitled to a meal allowance for a three-hour layover?”</em> HR steps in only for exceptions or appeals. The role of HR shifts from doing the work to validating AI outputs and adding judgment.</p><h3>The Next 3–5 Years: Insights and Collaboration</h3><p>Deloitte’s pilots show AI can forecast skill shortages. Imagine HR in corporate travel using AI to predict burnout by combining signals from overtime, project loads, client satisfaction, and travel intensity. AI provides the <em>what</em> and the <em>when.</em> HR provides the <em>why</em> and the <em>how.</em></p><p>If AI predicts burnout in Tokyo, it might suggest reducing hours. HR could instead rotate high-pressure client accounts, or offer flexible arrangements. But this only works if leaders are involved. Data alone is not enough. Leaders must collaborate with HR, discuss the findings, and act on them with empathy.</p><p>Today much of HR data is still manual. In the coming years, AI will help mine it faster. Surveys will still be needed to take the pulse of employees. AI accelerates analysis, but trust comes from HR listening, interpreting, and responding thoughtfully.</p><h3>Beyond 5 Years: Custodians of Fairness</h3><p>As economic pressures rise, AI may recommend blunt cost-cutting, laying off 20 percent of staff. The math may be right. But HR must remain the moral compass, weighing redeployment, retraining, and hybrid approaches. Firms like EY and KPMG are already experimenting with AI simulations of workforce futures, but they emphasise that HR must be the custodian of fairness, ensuring transitions protect dignity as much as profitability.</p><h3>Leadership, Data, and Collaboration</h3><p>AI cannot fix culture. Leaders can. The key is shared ownership of data. HR and leadership must work together on dashboards, combining surveys with AI insights. When leaders are engaged early, employees see that data is a dialogue, not a verdict. That is how psychological safety is built, and empowerment made real.</p><h3>Building the Data Foundations</h3><p>Governance may sound tedious, but it is the foundation of good AI. If policies are scattered, grievance records incomplete, or hours data missing, AI will misfire. HR needs clean schemas, clear ownership, simple rules, and handbooks that explain how data will be used.</p><p>With those basics in place, AI can validate and accelerate. Without them, AI becomes noise.</p><h3>Governance and Legal Oversight in Singapore</h3><p>AI also operates within the law. In Singapore, the <strong>Workplace Fairness Act (WFA)</strong> and the <strong>Personal Data Protection Act (PDPA)</strong> now set the standard. Together, they raise expectations. Grievances must be handled fairly and consistently, with privacy respected. Documentation is both a compliance requirement and a protection against disputes.</p><p>This means HR must carefully map access to grievance information, anonymise details where possible, and maintain role-based access. Investigations must balance confidentiality with fairness, ensuring employees accused of misconduct have enough information to respond.</p><p>AI may help draft templates or summarise trends. But HR ensures compliance, checks for bias, and trains managers to apply processes fairly.</p><p>As one counsel explained,<em>“The WFA and the PDPA are two sides of the same coin. The key is embedding fairness and privacy into every stage of grievance handling.”</em></p><blockquote>AI can speed the work. HR protects the trust.</blockquote><h3>What Skills Will Endure</h3><p>Even as AI grows, certain skills remain beyond automation. Emotional intelligence remains central given that no machine consoles an exhausted employee after a 14-hour shift. Ethical judgment will be critical, as laws like the WFA or EU’s AI Act demand human oversight of decisions. Strategic foresight will matter and predictions are useful only if HR can interpret cultural and organisational causes. And change management will be indispensable, as HR helps employees adapt to AI without losing their sense of value.</p><p>AI may draft communications or propose engagement ideas. HR validates them against law, culture, and fairness.</p><h3>HR Reframed, Not Replaced</h3><p>AI is rewriting the rulebook for HR. Budgets are shrinking, operations are streamlined, and some decisions are now machine-led. But the essence of HR that includes building trust, guiding leaders, protecting culture cannot be replaced.</p><p>The path forward requires courage. HR must experiment with AI tools, govern them to prevent bias, and reskill teams to combine technological fluency with human insight.</p><p>If HR reduces itself to efficiency, it risks irrelevance. But if it takes its place as custodian of culture and conscience of AI, HR will not only survive this transformation but lead it.</p><blockquote>The future of HR is not to compete with AI. It is to guide it ensuring technology serves humanity, not the other way around.</blockquote><h3>References</h3><ul><li>Axios. (2025). <em>Managerial use of AI</em>. Axios. <a href="https://www.axios.com/">https://www.axios.com</a></li><li>BusinessWire, Concur, &amp; Boston Consulting Group. (2024–2025). <em>Corporate travel AI adoption studies</em>. BusinessWire. <a href="https://www.businesswire.com/">https://www.businesswire.com</a></li><li>Deloitte. (2024–2025). <em>Human capital trends and workforce evolution surveys</em>. Deloitte Insights. <a href="https://www2.deloitte.com/">https://www2.deloitte.com</a></li><li>Gartner. (2025). <em>AI in HR adoption trends</em>. Gartner Research. <a href="https://www.gartner.com/">https://www.gartner.com</a></li><li>HireBee. (2025). <em>AI recruitment statistics</em>. HireBee. <a href="https://www.hirebee.ai/">https://www.hirebee.ai</a></li><li>IBM. (2024). <em>HR budget reduction through automation</em>. IBM Research. <a href="https://www.ibm.com/">https://www.ibm.com</a></li><li>PeopleSpheres. (2025). <em>HR automation report</em>. PeopleSpheres. <a href="https://www.peoplespheres.com/">https://www.peoplespheres.com</a></li><li>PwC. (2024). <em>Responsible AI survey</em>. PwC Global. <a href="https://www.pwc.com/">https://www.pwc.com</a></li><li>Tulane Law Blog. (2024). <em>AI ethics in HR</em>. Tulane University Law Blog. <a href="https://law.tulane.edu/">https://law.tulane.edu</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7b1808b94ea5" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Humanizing AI]]></title>
            <link>https://medium.com/@prabu.sangar/humanizing-ai-fb753ba01687?source=rss-ed44d3335f1d------2</link>
            <guid isPermaLink="false">https://medium.com/p/fb753ba01687</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[responsible-ai]]></category>
            <category><![CDATA[future-of-work]]></category>
            <category><![CDATA[technology-and-society]]></category>
            <dc:creator><![CDATA[Prabu Sangar]]></dc:creator>
            <pubDate>Sun, 07 Sep 2025 08:15:34 GMT</pubDate>
            <atom:updated>2025-09-07T08:15:34.699Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*q5Edj2Sl8hzCcBKE" /><figcaption>Photo by <a href="https://unsplash.com/@omilaev?utm_source=medium&amp;utm_medium=referral">Igor Omilaev</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h3>The Quiet Tax You Recognize Instantly</h3><p>It was 11:40 p.m. and the sixth prompt finally produced, “I understand how you feel”. The reply was polite and on point. The task that mattered still was not done. That is the quiet tax of humanizing AI.</p><p>We sand synthetic language until it feels warm, then realize the original work could have been completed faster by a clear document, a simple form, or a five‑minute human reply. Maya Angelou said, “People will never forget how you made them feel”. The feeling matters. The work still needs doing.</p><h3>Teaching A Model To Write KPIs</h3><p>Here is the moment that made this real for me. I set out to train ChatGPT to draft <strong>SMART </strong>KPI goals for individual staff. The idea sounded efficient.</p><p>I created a structured prompt, a scoring rubric, example goals from past years, and guidance on tone so the output read like our organisation. For a while it looked promising. Then the friction appeared. The model could not see baseline performance, stretch tolerance, upstream dependencies, or the very human politics of measurement. It handled wording, but it did not understand trade‑offs. After several hours of coaching and tests, I wrote the KPIs myself in less time. The lesson was not that the model is weak.</p><blockquote>The lesson is that some tasks are primarily judgment tasks disguised as writing tasks.</blockquote><h3>Adoption Is High, Impact Is Uneven</h3><p>According to the Stanford AI Index, reported AI incidents have continued to rise even as adoption accelerates, a reminder that responsible use rarely keeps perfect pace with deployment. According to McKinsey’s latest global survey, more than seven in ten organizations now use generative AI in at least one function, yet most leaders do not report enterprise‑level financial impact. The same research shows teams are still rewiring workflows and deciding how much human review is necessary. The excitement is genuine, the value is real in the right places, and the maturity curve is uneven.</p><p>A large field study by Erik Brynjolfsson and colleagues found that giving customer‑support agents an AI assistant increased issues resolved per hour by about fifteen percent on average, with the biggest gains for less experienced agents and little change for the most experienced. That is a useful map. It says AI can accelerate the basics and raise the floor, while expert judgment still sets the ceiling.</p><h3>What “Humanizing AI” Actually Means</h3><p>Humanizing AI is not a vibe. It is concrete work that includes tone guidelines, guardrails, refusal patterns, safe phrases, and escalation rules so a model sounds appropriate and does not harm people. The effort is necessary in sensitive domains, but the time is easy to underestimate.</p><p>W. Edwards Deming put it plainly, “In God we trust, all others must bring data”. If you do not measure the hours you spend tuning tone, you will confuse a nicer‑sounding bot with a better‑performing system.</p><h3>AI Helps Or Hurts</h3><p>Every AI system runs on two clocks.</p><ul><li><strong>Time to task</strong> is how quickly it produces a correct result.</li><li><strong>Time to trust</strong> is how long people take to accept and use that result.</li></ul><p>Most humanizing work moves the trust clock. If the task clock is slow or the answers are shaky, warmth will not save the experience. Peter Drucker’s reminder applies here, “What gets measured gets managed”. Define success as resolution with evidence, not as a generic sense of empathy.</p><p>You have seen both clocks at work. The service desk adds a friendly assistant that greets you by name, mirrors your sentiment, and apologizes gracefully. If the bot still fails to identify the right fix, you will not trust the next answer, no matter how kind it sounds. The reverse is also familiar. A plain tool that cites the exact policy clause and shows a short calculation earns trust quickly, even if it speaks like a manual.</p><h3>Reasoning With AI Is Hard</h3><p>Reasoning with AI is difficult because language models do not think in the human sense. They predict the next likely word given the words before it. That makes them excellent at pattern and style, and unreliable when a task demands explicit steps, stable world models, or private context. The model does not know what your team already tried, which trade‑offs your leadership has accepted, or where the political guardrails really sit. Ask it to “reason” about ambiguous constraints and you often get confident prose that masks uncertainty.</p><p>There are structural limits too. Memory is short. Retrieval can be partial. Tool use can fail quietly. When you chain multiple steps, small errors compound and the output may still sound authoritative. That is why a gentle sentence can be dangerously persuasive.</p><p>Alfred Korzybski’s line belongs here, “The map is not the territory”. The output is a map of language, not the territory of your organisation. George Box offered the companion truth, “All models are wrong, but some are useful”. Treat language models as useful when they stay close to evidence and tools, and wrong when you demand judgment they cannot possibly hold.</p><h3>Media Narratives Versus Operational Reality</h3><p>The media tells a hopeful story. It shows AI composing symphonies, drafting strategy, and acting as an ever‑present advisor. Vendors tell an adjacent story. They show agents orchestrating whole workflows that look like reasoning. In reality, most success today comes from well‑scoped tasks with strong ground truth, summarizing a set of documents that you supply, drafting a message from an approved template, extracting entities from a record, or running a calculation with guardrails and tests. The gap is not a scandal. It is a maturity gap.</p><p>You feel this gap when a keynote promises “AI that reasons about your business”, and you return to an environment where your policy is half in email, half in SharePoint, and the rest in people’s heads. The model does not fail because AI is overhyped. It fails because reasoning requires structure. If the facts are scattered, the agent will guess. When companies fix the plumbing first, the story improves. A retrieval index with current policies, a few verified tools, a short checklist for outputs, and clear escalation rules will outperform a charming agent with nothing firm to stand on.</p><h3>The Places We Spend Time That Do Not Pay Us Back</h3><p>Teams lose weeks to indefinite tuning because the goal is vague. If success is “sounds like us”, the work never ends. If success is “resolves the query in one step with a verifiable source”, tuning stops when the metric is met.</p><p>Another trap is outsourced empathy. A machine can simulate care, but staff can tell when a message is engineered to sound caring rather than written by someone who cares. That gap widens distrust. You do not rebuild trust with placeholders. You rebuild it with presence.</p><h3>Where Humanizing Is Non‑Negotiable</h3><p>There are moments when warmth is duty, not decoration. An apology to a customer after a failure. A message to staff during a restructure. A safety instruction that carries emotional weight. In these moments the fastest path is a thin, human‑written core that tells the truth clearly, followed by AI that formats and localizes. The human decides what to say. The model helps with how to say it. Albert Einstein warned, “Everything should be made as simple as possible, but not simpler”.</p><h3>Design In Layers</h3><p>There is a practical pattern that keeps teams out of the weeds.</p><ol><li>Start with a plain, verifiable core that does the work.</li><li>Retrieve policy and cite the paragraph.</li><li>Show the steps of a calculation.</li><li>Explain recommendations in one or two sentences that expose the reasoning.</li></ol><p>Put this core in front of real users and confirm it is correct, short, and useful. Only after the core earns trust should you add a light tone layer at the edges. Keep the layer thin and honest. Escalate to a human when the system detects distress, confusion, or conflict. When you design this way, you discover that a little humanizing goes a long way because utility has already done most of the trust building.</p><h3>Why Not SMART KPIs</h3><p>SMART goals look like structured text, which tempts teams to automate them. In reality, they are structured choices. They live inside baselines, interdependencies, seasonality, and local constraints. They must also feel fair to the person who carries them.</p><p>Research on performance management notes that managers already spend around two hundred hours a year on reviews, much of it on calibration and translation of strategy into individual outcomes. That time is not wasted bureaucracy. It is judgment. A model can help with phrasing, examples, and checks for measurability. It cannot own the trade‑offs or the politics of what the organisation chooses to reward.</p><h3>Everyday Scenes You Will Recognize</h3><p>The HR helpdesk launches a personable assistant that says “I am here to help” in five languages. Employees still struggle because the policy text is vague. A one‑page plain English policy and a short video would have solved the confusion by Monday. The marketing team tunes an AI to write warm emails that reference past purchases. Open rates do not move until the offer is clearer and the subject line is shorter.</p><p>The travel operations group experiments with an agent that greets consultants by name and ends with a polite sign‑off. The greetings are lovely. What the team actually needs is an agent that extracts fare rules, flags ticketing risks, and proposes a one‑line recommendation in the queue. Utility builds trust faster than a synthetic smile.</p><h3>What To Do Now, Without The Theatre</h3><p>Decide the problem before you decide the persona. If the problem is speed and accuracy, humanizing can wait. If the problem is fear or confusion, a person should speak. If the problem is both, pair a tool that does the work with a human who stands behind it. Measure both clocks. Track how long the system takes to produce a correct answer and how long it takes to make that answer acceptable to users. Change the design, not only the tone.</p><h3>The Call To Action</h3><p>Audit one workflow this week that you have been trying to humanize. Strip it to the smallest useful core and ship that version to a real user. Time how long it takes to deliver a correct answer, and time how long it takes to adjust tone. If humanizing consistently costs more time than it saves, cap it and move the budget to clarity, evidence, and escalation paths. For goals and reviews, let AI brainstorm examples and tighten phrasing, but let a manager write the actual SMART targets and own the trade‑offs.</p><p>The future belongs to teams that make AI useful before they make it charming. When charm matters, make it human. When you hear the promise that AI can reason about your business, ask an Eisenhower question, is the plan useful because it is perfect, or because it forces you to do the hard thinking now. “Plans are worthless but planning is everything”.</p><h3>References</h3><ol><li>Bender, E. M., Gebru, T., McMillan‑Major, A., &amp; Shmitchell, S. (2021). On the dangers of stochastic parrots: Can language models be too big. <em>Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency</em>, 610–623.</li><li>Box, G. E. P. (1976). Science and statistics. <em>Journal of the American Statistical Association, 71</em>(356), 791–799.</li><li>Brynjolfsson, E., Li, D., &amp; Raymond, L. (2023). <em>Generative AI at work</em> (NBER Working Paper №31161). National Bureau of Economic Research.</li><li>Drucker, P. F. (1973). <em>Management: Tasks, responsibilities, practices</em>. Harper &amp; Row.</li><li>McKinsey &amp; Company. (2025). <em>The state of AI: How organizations are rewiring to capture value</em>.</li><li>Stanford Institute for Human‑Centered AI. (2025). <em>AI Index 2025: Policy highlights</em>.</li><li>Society for Human Resource Management. (2019). <em>The performance review process can be a lot easier. Here’s how</em>.</li><li>Deming, W. E. (1986). <em>Out of the crisis</em>. MIT Press.</li><li>Kahneman, D. (2011). <em>Thinking, fast and slow</em>. Farrar, Straus and Giroux.</li><li>Korzybski, A. (1933). <em>Science and sanity: An introduction to non‑Aristotelian systems and general semantics</em>. Institute of General Semantics.</li><li>Einstein, A. (1950). <em>Out of my later years</em>. Philosophical Library.</li><li>Eisenhower, D. D. (1962). Remarks at the National Defense Executive Reserve Conference. <em>Public Papers of the Presidents of the United States</em>.</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fb753ba01687" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>