<?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 Arun Ragothaman on Medium]]></title>
        <description><![CDATA[Stories by Arun Ragothaman on Medium]]></description>
        <link>https://medium.com/@arunragothaman?source=rss-abb2484838d4------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*oMpJIO1Xo-Ykf9-rHojyBQ.jpeg</url>
            <title>Stories by Arun Ragothaman on Medium</title>
            <link>https://medium.com/@arunragothaman?source=rss-abb2484838d4------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 01:57:45 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@arunragothaman/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[Why your AI programme stalled. And it wasn’t the technology]]></title>
            <link>https://medium.com/@arunragothaman/why-your-ai-programme-stalled-and-it-wasnt-the-technology-18cf277ca743?source=rss-abb2484838d4------2</link>
            <guid isPermaLink="false">https://medium.com/p/18cf277ca743</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[banking]]></category>
            <category><![CDATA[ai-adoption]]></category>
            <category><![CDATA[ai-agent]]></category>
            <category><![CDATA[enterprise-technology]]></category>
            <dc:creator><![CDATA[Arun Ragothaman]]></dc:creator>
            <pubDate>Thu, 04 Dec 2025 11:30:39 GMT</pubDate>
            <atom:updated>2025-12-04T11:30:39.623Z</atom:updated>
            <content:encoded><![CDATA[<p>The business case was bulletproof. Nine months of rigorous review, every assumption tested, every saving validated. Another three months mapping benefits to the right ledger codes. Sixteen months from proposal to signed contract — thorough, methodical, unassailable.</p><p>Four months later, a reorganisation shifted reporting lines and merged divisions. The headcount assumptions no longer held. The AI programme hadn’t failed; it had simply been orphaned by an organisation that moved around it.</p><p>This wasn’t a technology problem. The platform worked. The models performed. But somewhere between signature and deployment, the ground shifted.</p><p>I’ve spent the past two years embedded in large-scale AI-led automation in banking. Close enough to operations to understand how work actually happens; present in enough boardrooms to see strategic intent. After three major programmes, certain patterns emerge. They have nothing to do with model architecture, and everything to do with how organisations absorb change.</p><p><strong>The meeting that changes when the sponsor attends (or leaves)</strong></p><p>Certain meetings are well attended when the executive sponsor is present. In their absence, the same agenda takes a different turn. One team pushes ahead. Another returns to first principles, questioning why the organisation is pursuing this at all. A third invokes past failures: the 2019 initiative that stalled, the 2021 pilot that never scaled. The meeting ends without decisions.</p><p>Oddly, the team working hardest to make the programme succeed is often the one whose function will be automated. They understand what’s at stake. Meanwhile, teams whose influence depends on controlling information flows discover urgent concerns about data governance, security, or compliance. AI programmes surface power structures that were previously implicit. Those structures defend themselves.</p><p><strong>The incentives that guarantee failure</strong></p><p>An executive measured on cost reduction champions the programme. The implementation team reports to a leader measured on headcount. Innovation mandates conflict with operational-stability bonuses. Business cases promise efficiency gains, but no one’s objectives actually require realising them.</p><p>This isn’t malice; it’s mechanics. When the innovation team is rewarded for launching pilots and the operations team penalised for disruption, the system produces exactly what you’d expect: successful demonstrations that never reach production.</p><p><strong>The process nobody fully understands</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cl1l_XP7wPAJ9Kcex5BF_g.png" /></figure><p>A workflow requires sixteen approval steps. Embedded within it are “temporary” workarounds that have operated for six years, undocumented. When the AI programme attempts to automate parts of this process, it emerges that only three people understand how it works. One is retiring.</p><p>The word “integration” triggers primal instincts. It requires at least two departments to coordinate, which means navigating unwritten rules about who defers to whom, whose budget bears the cost, and whose roadmap takes priority. Data is technically accessible ; the APIs exist, permissions can be granted, but it’s practically locked behind gatekeepers who’ve built careers on knowing where things are.</p><p>The gap between how work is supposed to flow and how it actually flows becomes visible only when you try to change it.</p><p><strong>What works</strong></p><p>The programmes that transform share common traits. They’re led by executives who attend steering committees, make decisions when obstacles emerge, and visibly remove barriers rather than endorsing initiatives from a distance. They recalibrate incentives before launching pilots, ensuring that those responsible for implementation benefit from success rather than absorb risk for disrupting business as usual.</p><p>These organisations acknowledge the political dimension openly. They don’t pretend that automating a department’s core function is neutral, or that integrating systems across divisions is merely technical. They create space for honest conversation about what changes and who’s affected. And they invest time mapping how work actually happens — not how the org chart suggests it should — because that’s where implementations either gain traction or quietly stall.</p><p><strong>The real bottleneck</strong></p><p>The technology challenges in AI adoption are solvable. Models improve monthly. Infrastructure matures. Vendors compete to simplify implementation. <em>Organisational dynamics don’t debug themselves.</em></p><p>The hardest problems aren’t in the algorithm or the vector database. They’re in the meeting that loses quorum when the sponsor travels. They’re in the performance template that rewards activity over outcomes. They’re in the unwritten rule about which departments collaborate and which compete.</p><p>If your AI programme has flawless technical execution but can’t clear the steering committee, you don’t have a technology problem. You have an organisation that hasn’t yet decided whether it wants to change.</p><p>The models are ready. <em>The question is whether the institution is.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=18cf277ca743" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Achieving Greater Accuracy in Sales Forecasts]]></title>
            <link>https://medium.com/@arunragothaman/achieving-greater-accuracy-in-sales-forecasts-5f5107dab44d?source=rss-abb2484838d4------2</link>
            <guid isPermaLink="false">https://medium.com/p/5f5107dab44d</guid>
            <category><![CDATA[sales]]></category>
            <category><![CDATA[b2b]]></category>
            <category><![CDATA[saas]]></category>
            <category><![CDATA[sales-forecasting]]></category>
            <dc:creator><![CDATA[Arun Ragothaman]]></dc:creator>
            <pubDate>Sun, 27 Apr 2025 03:15:22 GMT</pubDate>
            <atom:updated>2025-04-27T03:15:22.153Z</atom:updated>
            <content:encoded><![CDATA[<h4><a href="https://pf.anaplan.com/finance-transformation/research-finance-leaders-survey">99% of organisations have experienced negative consequences due to decisions based on inaccurate forecasts.</a></h4><h3><strong><em>Achieving Greater Accuracy in Bottom-Up Sales Forecasts</em></strong></h3><blockquote><em>“When the grass is dry at night, look for rain before the light”,<br>“When the grass is dry at morning light, look for rain before the night.”</em></blockquote><p>A hundred years ago, boy scouts used these tips to predict weather. Today, we have phones for accurate weather forecasts, and don’t need to go looking for pastures. Yet, when it comes to sales forecasting, we rely on feelings over facts.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2Ciw6bAGhTztT7vu-3qrYw.jpeg" /><figcaption>Image credit: <a href="https://www.instagram.com/somestupidcat/">@somestupidcat</a> / <a href="https://www.linkedin.com/in/teresa-wong-b6221584/">Teresa Wong</a></figcaption></figure><p>Even with technological advances, sales forecasting for complex cycles is still unreliable. Before exploring why, let’s understand what forecasts are and why they matter.</p><p>Forecasts estimate future revenue much as sailors predict conditions before setting sail. Accurate forecasts grow businesses by guiding better decisions and enabling efficient resource allocation. Conversely, inaccurate forecasts put a strain on leadership, salespeople, and customer relationships. In some cases, they lead to business closures.</p><p>Consequences of inaccurate forecasts:</p><ol><li>Leadership loses trust with the board when they cannot allocate resources effectively or set accurate sales targets. They dwell on problems and focus less on growth.</li><li>Salespeople’s performance suffers when targets are unrealistic. A last-minute rush to close deals often leads to poor agreements that either don’t last or fail to generate profit. Top salespeople perform best with clear, realistic goals; without them, companies risk losing their best talent, slowing growth.</li><li>Customers lose trust when companies repeatedly fail to meet commitments due to ineffective resource allocation. <a href="https://www.economist.com/business/2018/03/28/getting-a-handle-on-a-scandal">The Economist magazine’s analysis of Volkswagen, Wells Fargo, and six other corporations found that companies lose 30% of their value — at least in the short term — when they lose trust.</a></li></ol><p><strong>Forecasts are important — but why do companies get them wrong?</strong></p><p>Businesses that supply complex solutions to other businesses, face tough challenges: long and unpredictable sales cycles, limited and often unclear information, and a heavy reliance on a few large deals to drive results. In these situations, the salespeople closest to the process are best placed to predict the chances of closing a deal.</p><p>It’s like sailing into uncharted waters without satellite data or historical records — relying solely on the experience of sailors who have navigated similar routes. Similarly, forecasts lean on salespeople’s judgement, which can be subjective and prone to error. This bottom-up approach is practical and convenient, yet unreliable.</p><p><a href="https://www.gartner.com/en/newsroom/press-releases/2020-02-12-gartner-says-less-than-50--of-sales-leaders-and-selle">According to Gartner, a research firm, the majority of sales leaders and sellers lack confidence in their forecasts</a>.</p><p>To understand why, we need to examine how organisational systems are designed and how human input is influenced by bias.</p><p><strong><em>One-size-fits-all fits no one.</em></strong></p><p>Every company aims to operate at scale. A unified customer relationship management (CRM) system and shared sales processes support scalability — making it easier to sell and serve different client segments. However, this creates challenges for teams.</p><p>Clients have different buying processes and fiscal calendars, yet the entire sales force is expected to follow the same process and forecast rhythm — usually weekly — across the organisation. Sellers and their managers must adapt the process to their specific territories, but this mismatch is both fundamental and permanent. Sellers are required to forecast weekly, even if their territory generates revenue only once or twice a year. It’s like using the same Ikea toolkit to build a chair, a boat, and a road.</p><p>Sales cycles, sellers, and clients maybe treated as exceptions, but when the exceptions outweigh the norm, the norms themselves may no longer be effective.</p><p><strong><em>Subjective inputs are, well, subjective.</em></strong></p><p>Both, Sales managers and sellers agree that forecasts are important. However, differing motivations produce biased results.</p><p>Seller’s Motivation</p><ul><li>Sellers are rewarded for closing deals, not for providing accurate forecasts.</li><li>A missed forecast usually results in just an awkward conversation — something most sellers manage with ease.</li><li>They often offer brief inputs to finish forecast calls quickly.</li></ul><p>Manager’s Challenges</p><ul><li>Limited visibility: Managers often lack useful insights into deals and have little autonomy to drive their own processes.</li><li>Dependence on skill: The quality of a forecast conversation relies heavily on the sales manager’s competence. If a manager is nervous or inexperienced, the forecast conversation can lose focus and become unreliable.</li></ul><p>The final forecast will be optimistic or pessimistic, but not accurate.</p><p><strong>How Do Companies Get Forecasting Right?</strong></p><p>Companies must upgrade systems to adapt artificial intelligence (AI) technology and empower sales managers to improve forecast accuracy.</p><p><em>Upgrade and tailor systems for distinct teams and clients.</em></p><p>Every company’s mission statement focuses on the customer. It is time to tailor systems and processes for them.</p><ol><li>Depending on the customer segment — mid-size, government, or large multinationals — systems should understand sales cycles, deal sizes, key decision-makers, deal communication status, and financial windows for closing deals. These insights must be surfaced to sales managers.</li><li>Sales processes can be tailored to the customer segment and embedded into systems, making forecast frequency congruous to length of deal cycle. For an 18-month strategic pursuit, there is little value in a weekly update to the CRM. However, for a mid-size territory with quick sales cycles, there are multiple updates that need to be captured in a week.</li></ol><p>These changes will lay the foundation for improved conversations between sellers and sales managers.</p><p><em>Trust creates candid conversations.</em></p><p>Organisations must empower sales managers, so they lead and own the forecast process.</p><ol><li>Sales managers should bring data-led inputs and ideas to the conversation — reversing the traditional forecast calls. They should avoid asking subjective questions about deals, as this encourages storytelling. <em>For example: This client usually makes a purchase in Mar or September. What demand can we create for 4 months from now? The system shows reduced communication in the past week. I have some ideas that I would like to share with you.</em></li><li>Sellers are motivated by incentives. Reward sellers who are factual and reflect reality. This would create an environment where it is easy to share “bad news”.</li><li>Sales managers should keep sales performance separate from forecasts. They should avoid including non-sellers in forecast calls, as it discourages candour.</li></ol><p>Until AI can fully automate complex B2B sales forecasts, human judgement will remain vital. By refining our systems, incentives, and conversations, we can transform forecasting from guesswork into a strategic advantage.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5f5107dab44d" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>