<?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 Greg Lakloufi on Medium]]></title>
        <description><![CDATA[Stories by Greg Lakloufi on Medium]]></description>
        <link>https://medium.com/@GregLakloufi?source=rss-2fa4496fe40b------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*uCXmUUPq-dnnc_X-svPbMA.jpeg</url>
            <title>Stories by Greg Lakloufi on Medium</title>
            <link>https://medium.com/@GregLakloufi?source=rss-2fa4496fe40b------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 23:21:20 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@GregLakloufi/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[The Organizations That Think Will Win.]]></title>
            <link>https://medium.com/@GregLakloufi/the-organizations-that-think-will-win-e1cd9fea045c?source=rss-2fa4496fe40b------2</link>
            <guid isPermaLink="false">https://medium.com/p/e1cd9fea045c</guid>
            <category><![CDATA[strategic-design]]></category>
            <category><![CDATA[systems-thinking]]></category>
            <category><![CDATA[business-strategy]]></category>
            <category><![CDATA[service-design]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Greg Lakloufi]]></dc:creator>
            <pubDate>Tue, 06 Oct 2026 19:53:27 GMT</pubDate>
            <atom:updated>2026-10-06T20:02:42.344Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AX4mV5W0GCWCWqCBME4ngw.png" /><figcaption>A future shaped not by smarter technology, but by smarter understanding.</figcaption></figure><h4>Why understanding is becoming the most valuable enterprise capability.</h4><p>There is another quiet shift taking place inside modern business. Unlike previous transformations, this one is unlikely to be remembered for a breakthrough technology, a disruptive startup, or a revolutionary management philosophy. It will not begin with a keynote presentation or a bestselling business book. There will be no single moment that historians point to and declare, “This is where everything changed.”</p><p>Like many of the most consequential shifts in business, it is already underway long before most of us have given it a name.</p><p>For much of the last century, competitive advantage followed a remarkably consistent pattern. Companies that designed, manufactured, and distributed better products tended to outperform those that did not. As products became easier to replicate and global competition intensified, advantage gradually shifted toward services and experiences. Companies that understood customers more deeply and orchestrated more meaningful relationships created value that competitors found increasingly difficult to copy.</p><p>That transition reshaped entire industries. It gave rise to disciplines such as User Experience, Customer Experience, and Service Design. It encouraged businesses to think beyond products and toward the broader ecosystems in which customers lived, worked, and interacted. The shift was profound because it reflected a larger truth: value had moved from what companies produced to how people experienced what they produced.</p><p>Today, another transition appears to be underway.</p><p>Artificial intelligence is transforming nearly every aspect of modern business. Companies are redesigning products, automating operations, accelerating software development, personalizing customer experiences, and reimagining entire business models around AI. Every executive agenda includes it. Every board meeting discusses it. Every consulting firm has built a practice around it. The technology is extraordinary, and its long-term impact may rival that of the internet itself.</p><p>Five years ago, access to frontier AI models was itself a competitive advantage. Today, companies can choose between multiple world-class models, each becoming more capable and more accessible with every passing quarter. Just as cloud computing gradually shifted from strategic differentiator to business infrastructure, artificial intelligence appears to be following a similar trajectory. The competitive question is already beginning to change from “Who has AI?” to “Who consistently makes better decisions <strong>because</strong> AI exists?”</p><p>Yet as significant as AI may become, I do not believe it represents the next enduring source of competitive advantage. What begins as a differentiator gradually becomes infrastructure. Electricity followed that pattern. The internet followed that pattern. Cloud computing followed that pattern. Artificial intelligence appears likely to follow a similar pattern.</p><p>If every organization eventually has access to increasingly capable AI, then advantage must once again move somewhere else.</p><p>The question is where.</p><p>The more I thought about that question, the more I realized it seemed to be hiding beneath almost every executive conversation I was having. Once I noticed it, I started seeing it everywhere. <strong>Leaders are not struggling to access information. They are struggling to make sense of it.</strong></p><p>Never before have enterprises invested so heavily in becoming more informed. They subscribe to analyst research, commission customer studies, monitor competitors, build strategy teams, deploy dashboards, and use artificial intelligence to synthesize enormous volumes of content. They have more data, expertise, and sophisticated tools than any previous generation of business leaders.</p><p>Modern enterprises have never known more. But it has rarely been harder to decide what to do next.</p><p>I couldn’t quite reconcile that contradiction. Then something unusual caught my attention. Within days of one another, Netflix and OpenAI published executive roles that seemed, at first glance, completely unrelated yet appeared to be solving the same organizational problem. The more I thought about those roles, the less interested I became in who they were hiring and the more interested I became in the problem they were trying to solve.</p><h3><strong>The Intelligence Paradox</strong></h3><p>One of the defining characteristics of modern enterprises is that they have become extraordinarily good at collecting intelligence. Over the past two decades, businesses have invested billions of dollars to better understand the world around them. Customer research, product analytics, competitive intelligence, business intelligence, strategy, social listening, forecasting, and now AI each generate their own stream of insight, all intended to help leaders make better decisions.</p><p>Viewed individually, these investments have been remarkably successful. Companies understand customer behavior with extraordinary precision. They can monitor competitors in real time, detect emerging market trends, forecast demand, and measure operational performance down to the smallest detail. Information that once took months to gather can now be surfaced almost instantly.</p><p>By every conventional measure, modern enterprises should be operating with unprecedented clarity. Yet the conversations inside executive boardrooms often tell a different story. The difficulty is no longer accessing information. It is determining which information deserves attention, how seemingly unrelated signals influence one another, and what actions leadership should take as a result.</p><p>Consider a large healthcare system evaluating whether to expand AI-assisted clinical documentation. The Chief Medical Officer sees physician burnout. Finance sees cost reduction. Compliance sees regulatory uncertainty. Technology sees integration challenges. Clinical leadership worries about trust. Patient experience teams see opportunities to improve care. Every perspective is valid. None independently answers the question executives are actually trying to solve.</p><p>Customer research recommends one course of action. Market analysts suggest another. Strategy identifies a different opportunity. Sales hears something else from clients. Product sees another priority. Finance introduces constraints. Legal raises regulatory concerns. AI can summarize every perspective, but the executive team still leaves the meeting without a shared understanding of what matters most.</p><p>In many cases, the opposite is true. Every function has become exceptionally good at understanding its own piece of the world, while very few enterprises have developed an equally sophisticated capability for connecting those pieces into one coherent picture.</p><p>For decades, that was not a significant problem. Business complexity generally evolved at a pace companies could absorb. Markets shifted, competitors launched products, regulations changed, technologies advanced, and customer expectations evolved. Each development could largely be interpreted within the context of a specific function. Marketing understood markets. Product managed products. Finance allocated capital. Operations optimized execution. Strategy looked ahead.</p><p>Today, a single event can cascade across an enterprise in ways that were almost unimaginable a decade ago. A breakthrough in artificial intelligence is not simply a technology story. It becomes a product strategy discussion, a governance challenge, a workforce planning exercise, a cybersecurity consideration, a regulatory concern, a customer experience opportunity, and a competitive positioning question at the same time.</p><p>Complexity has become systemic. Most companies, however, continue to process it through structures designed for a simpler world.</p><p><a href="https://medium.com/@GregLakloufi/the-age-of-systems-why-products-and-services-are-no-longer-enough-ec77e50d947f">In a previous article, I argued that competitive advantage had quietly moved from products and services toward <strong>systems</strong>.</a> That observation naturally raises another question. If systems have become the new source of competitive advantage, what capability allows enterprises to continuously understand, shape, and adapt those systems before everyone else does? The more I reflected on that question, the more convinced I became that I was looking at the answer.</p><p>The problem is no longer discovering signals. <br><strong>The problem is connecting them.</strong></p><h3><strong>A Signal Hiding in Plain Sight</strong></h3><p>The first clue came from a job posting at <a href="https://www.linkedin.com/jobs/view/4438407883/?trackingId=6HFHwZj3Q0O%2BrAnOfaCKmw%3D%3D">Netflix</a>. A few days later, another appeared at <a href="https://www.linkedin.com/jobs/view/4436853932/?trackingId=aw0c60%2F0SOyJs7LbQ86DGA%3D%3D">OpenAI</a>. At first glance, there was nothing particularly remarkable about either role. Large companies hire senior leaders all the time. Executive titles come and go. Responsibilities evolve. New functions emerge. It is easy to dismiss them as ordinary hiring activity.</p><p>Then curiosity got the better of me. Instead of reading the title, I started reading the job description. That turned out to be where the real story was hiding.</p><p>One of the positions was called Head of Competitive Intelligence. On the surface, it sounded familiar. Competitive Intelligence has existed for decades. Most large enterprises maintain some capability to monitor competitors, analyze market movements, support sales teams, or evaluate strategic threats. There was nothing revolutionary about the title itself.</p><p>The responsibilities, however, described something significantly broader. The role was expected to connect market movements to product strategy, translate competitive dynamics into executive decision-making, identify emerging opportunities, shape investment priorities, influence pricing, support go-to-market strategy, monitor partnerships, interpret technological shifts, evaluate ecosystem changes, and help leadership understand how all of those forces interacted.</p><p>The further I read, the more I started wondering if I was looking at the wrong thing entirely. The title stopped matching the work. This was not traditional Competitive Intelligence. <strong>It was a mechanism for transforming external complexity into internal clarity.</strong></p><p>Imagine the executive conversation taking place inside those companies. An executive team concluded that existing functions were no longer enough. Market research alone was insufficient. So were customer research, product management, strategy, business intelligence, and artificial intelligence. Each understood an important part of the landscape. None, individually, could help leadership continuously make sense of all of it.</p><p>Companies do not create executive roles like these merely in response to trends. <strong>They create them in response to capability gaps.</strong> Netflix and OpenAI appear to have reached a similar conclusion from very different starting points: <strong>extraordinary intelligence already existed across the enterprise, but the connective mechanism required to turn that intelligence into coherent strategic action did not.</strong></p><p>When several of the world’s most sophisticated companies begin investing in the same capability at roughly the same time, they are rarely responding to yesterday’s problem. <strong>They are quietly preparing for tomorrow’s.</strong></p><p>What caught my attention wasn’t that Netflix and OpenAI were hiring similar roles. It was that two very different organizations had independently concluded they needed <strong>a new way to turn intelligence into action.</strong> I was looking at <strong>the emergence of a new organizational capability, one designed not to generate more information, but to continuously transform fragmented signals into shared understanding, strategic clarity, and better decisions.</strong></p><p>More recently, I noticed the same signal somewhere I did not expect to find it: the U.S. Intelligence Community. The <strong>Office of the Director of National Intelligence</strong> is recruiting a <a href="https://www.linkedin.com/jobs/view/4475568042/?refId=OivrQ%2BtVNQV3fBczBQM07Q%3D%3D&amp;trackingId=OivrQ%2BtVNQV3fBczBQM07Q%3D%3D">Chief Strategy Officer</a> whose mandate includes evaluating strategic trends, navigating cross-community interdependencies, aligning resources to strategy, assessing readiness for future environments, facilitating competing priorities, and turning strategy into implementation.</p><p>Netflix. OpenAI. National intelligence.</p><p>Very different missions. Very different environments. Yet each points toward the same emerging need: the ability to connect complexity into understanding and understanding into action.</p><p>History suggests something predictable happens whenever a new enterprise capability emerges. The first organizations build it. Everyone else eventually needs it.</p><h3>A Familiar Pattern</h3><p>Every meaningful shift in business follows a remarkably familiar path.</p><p>A changing environment exposes a problem that existing disciplines can no longer solve. A handful of pioneering companies recognize it first. They experiment. They create new roles, new operating models, and new ways of working. What begins as an isolated capability gradually becomes something much larger.</p><p>Over time, an ecosystem begins to form. Universities develop courses. Technology platforms emerge. Professional communities take shape. Standards evolve. New executive roles appear. Research expands. Advisory firms, software providers, and implementation partners all contribute to helping the capability mature. Eventually, what once felt experimental becomes an accepted part of how modern enterprises operate.</p><p>We have seen this pattern before.</p><p>Customer Experience evolved from a competitive differentiator into an established discipline. Cybersecurity followed a similar path, transforming from a technical concern into a board-level responsibility supported by an entire ecosystem of technologies, practices, standards, and expertise. Cloud engineering, enterprise architecture, and data science each followed remarkably similar trajectories.</p><p>Somewhere along the way, I realized I had been asking the wrong question. Instead of asking why Netflix or OpenAI are hiring these roles, the more interesting question becomes whether we are witnessing the early formation of another enterprise discipline.</p><p>Thirty years ago, very few companies employed Chief Information Security Officers. Today, cybersecurity is recognized as an essential organizational capability. Only the world’s largest enterprises build every aspect of that capability themselves. Most participate in a much broader ecosystem of technologies, academic research, professional communities, specialist expertise, and trusted partners that continues to evolve alongside the discipline itself. Organizational understanding may be following a remarkably similar path.</p><p>The complexity does not discriminate based on company size.<br>Only the ability to respond to it does.</p><p>Perhaps the real opportunity is not to create another Competitive Intelligence function or another market research service. It is to help enterprises develop a repeatable capability for transforming fragmented signals into coherent understanding and better decisions.</p><p>Information informs decisions.<br>Understanding shapes them.</p><p>If history is any guide, we may be witnessing something far more significant than the emergence of a new role. We may be watching the early formation of an entirely new enterprise capability. Whether it eventually becomes a discipline is almost beside the point.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SXahPwMcY8qWv7P51m_p5w.png" /><figcaption>The competitive advantage is no longer information. It’s the ability to connect it.</figcaption></figure><h3>When a New Discipline Emerges</h3><p>One implication became increasingly difficult to ignore. If this emerging capability is real, then only a very small number of organizations will build it internally. Companies like Netflix and OpenAI can justify hiring senior executives dedicated to organizational intelligence. They possess the scale, complexity, and resources to create entirely new enterprise functions.</p><p>Most do not. Yet they face remarkably similar challenges.</p><p>Healthcare providers still need to understand how AI, regulation, reimbursement, customer expectations, and workforce shortages interact. Manufacturers still need to connect geopolitical shifts, supply chains, automation, sustainability, and capital investment. Financial institutions still need to navigate regulatory complexity, emerging technologies, changing customer behavior, cybersecurity, and competitive disruption. The complexity is shared.</p><p>Only the capacity differs.</p><p>History suggests this is exactly how entirely new consulting categories emerge. The first industry leaders build a capability. The rest eventually seek partners capable of helping them build it too.</p><p>Not by outsourcing the thinking. By accelerating the organization’s ability to think for itself.</p><p>That distinction is important.</p><p>The goal is not to become another source of intelligence. The goal is to help clients develop their own intelligence.</p><p>To design the operating rhythms, decision structures, synthesis capabilities, executive rituals, governance models, and organizational behaviors that allow understanding to become a repeatable enterprise capability rather than an occasional executive exercise. The most valuable consulting firms of the next decade may not be those that deliver the best recommendations. They may be those that leave behind organizations capable of generating better decisions long after the consultants have gone.</p><p>Those that build this capability internally will create competitive advantage. The organizations that cannot will increasingly seek partners capable of helping them design it.</p><p>The question is no longer whether organizations will need this capability. It is whether they will build it themselves or seek partners to help them develop it. That realization led me somewhere unexpected. If organizations needed a new capability for making sense of complexity, I couldn’t shake the feeling that I had seen many of its underlying skills before. Just not where most people would think to look.</p><h3>Design Was Always Headed Here</h3><p>Looking back, I do not think this shift emerged from a single design discipline. It emerged through design’s gradual expansion of perspective. Industrial Design taught us to shape products. Interaction Design taught us to shape digital behavior. Experience Design taught us to shape journeys. Service Design taught us to shape systems. Each expanded the designer’s field of view. Each prepared us for what comes next.</p><p>I kept finding myself returning to Service Design. At first, I wasn’t entirely sure why. Frankly, that surprised me. It wasn’t because Service Designers have traditionally worked inside corporate strategy or owned competitive intelligence. Most haven’t. What kept drawing me back was something deeper than methods, deliverables, or organizational charts. It was the way the discipline teaches us to think.</p><p>Long before businesses began speaking about ecosystems, resilience, interconnectedness, or artificial intelligence, Service Design encouraged us to see relationships where others saw functions. It challenged us to look beyond the visible artifact and understand the invisible structures producing it. It asked us to move continuously between people and systems, business objectives and human outcomes, operational realities and customer experiences. Most importantly, it taught us that the visible experience was rarely the real problem. The real problem usually lived somewhere deeper inside the system.</p><p>Perhaps we’ve been underselling ourselves. For years, we have said that Service Design helps businesses design better services. Perhaps that description was always too small.</p><p>Perhaps services simply happened to be the systems we were asked to improve.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gVNqHYSclLu9QR_O_ZL20Q.png" /><figcaption>Every fragmented system eventually becomes someone’s lived experience.</figcaption></figure><p>The more I sat with that thought, the bigger it became.</p><p>Take a very human problem that seems almost painfully ordinary. Think about a patient trying to schedule oncology treatment after receiving a life-changing diagnosis. The frustration rarely begins with the appointment itself. It begins much earlier, inside disconnected scheduling systems, reimbursement policies, staffing shortages, referral processes, and organizational incentives invisible to the patient. Service Design taught us to look beneath the experience and understand the system producing it. The enterprises of tomorrow may need that same capability applied not only to customer experiences, but to strategic decision-making itself.</p><p>Journey maps, service blueprints, workshops, and touchpoints are expressions of a deeper capability: the ability to synthesize complexity, reveal hidden relationships, and make sense of systems that others experience only in fragments. The discipline was quietly developing a capability the business world is only now beginning to need.</p><p>Leadership teams are increasingly confronting questions that do not belong to a single function. How will advances in artificial intelligence reshape customer expectations? What happens to an operating model if regulation changes in Europe? How should a new competitor influence the product roadmap? Which partnerships strengthen an ecosystem, and which increase dependence? Where are the weak signals suggesting an entirely different market is beginning to emerge?</p><p>These questions exist between disciplines, and that is precisely where Service Design has always done its best work. The scarce capability is no longer simply the ability to gather or analyze information. It is the ability to connect: to recognize patterns before they become obvious and understand how technological breakthroughs, geopolitical events, regulation, workforce dynamics, and customer behavior combine to influence the decisions leaders make tomorrow morning.</p><p>That is not analysis. It is synthesis.</p><p>For years, many of us have described Service Designers as translators. We translate between customers and businesses, technology and operations, frontstage experiences and backstage realities. The longer I’ve sat with this idea, and the more I reflect on the challenges leaders now face, the less convinced I’ve become that “translator” is the right word.</p><p>Translators move information from one place to another.<br>Integrators create understanding.</p><p>An integrator helps leaders recognize patterns that no individual function can see on its own. They connect customer behavior with technological change, regulatory shifts with business strategy, geopolitical uncertainty with investment decisions, and emerging capabilities with human needs. Most importantly, they help leadership understand not only what is happening, but why it matters.</p><p>Perhaps design’s greatest contribution over the past two decades has not been teaching businesses how to create better experiences or better services. It has been teaching us how to think across systems before business had a language for doing so.</p><p>Service Design may have developed many of these capabilities first. The opportunity now is to extend them beyond services and apply them to how entire enterprises perceive, interpret, and act.</p><p>Which raises the obvious question: what exactly is the capability we have been building?</p><h3>Designing Understanding</h3><p>I’ve started wondering if we’ve been asking the wrong question. Every era seems to demand a new organizational capability. The Industrial Age rewarded operational excellence. As markets matured and products became increasingly interchangeable, advantage migrated toward customer experience. Today, artificial intelligence is accelerating the production of knowledge at an unprecedented rate, exposing an entirely different scarcity.</p><p>Information scales remarkably well.<br>Understanding does not.</p><p>Understanding emerges when seemingly unrelated signals are interpreted through context, challenged through multiple perspectives, and connected into narratives that help leaders decide where to move next. It depends as much on judgment as analysis, as much on synthesis as expertise, and as much on relationships as information.</p><p>The nature of executive decision-making has fundamentally changed. Consider a CEO deciding whether to make a major investment in artificial intelligence. A decade ago, this might have been viewed primarily as a technology decision. Today it is simultaneously a workforce, governance, cybersecurity, customer experience, regulatory, investment, product strategy, and geopolitical decision. Every perspective is valid. None, independently, is sufficient.</p><p>The quality of the decision no longer depends on the quality of any single analysis. It depends on the organization’s ability to connect them.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*9ChNXnQ0pRWpQAIFRV0ZsQ.png" /><figcaption>In contested environments, decision speed matters. But speed without understanding can simply accelerate the wrong decision.</figcaption></figure><p>Imagine, for a moment, a senior Space Force commander evaluating intelligence that suggests a strategic competitor may be preparing to disrupt critical satellite capabilities. Orbital analysts observe unusual spacecraft maneuvers. Intelligence agencies report increased military signaling. Cyber teams detect probing activity against ground infrastructure while commercial satellite operators notice unexpected launch patterns. Diplomats warn of rising geopolitical tension. None of these signals independently justify action. Together, they may reveal an entirely different reality.</p><p>Every expert has done exactly what they were trained to do. The challenge is that no single discipline is responsible for connecting what everyone knows before a decision can no longer wait.</p><p>Perhaps the next frontier is therefore not another dashboard, report, workshop, or strategic assessment. It is designing the conditions through which enterprises continuously transform fragmented signals into coherent understanding: how intelligence flows, how meaning emerges, how confidence develops, and how better decisions become possible.</p><p>Strategy, product management, customer research, business intelligence, competitive intelligence, finance, marketing, operations, and artificial intelligence all create extraordinary value. The challenge is that each has become exceptionally good at answering its own questions. Very few are designed to answer the questions that exist between them.</p><p>The capability emerging to answer those questions is not simply another evolution of strategy or research. It is the intentional design of organizational understanding.</p><p>I think of this emerging capability as Design Intelligence: the deliberate design of how organizations perceive, synthesize, interpret, and act upon an increasingly complex world. Its purpose is not to create more knowledge, but to orchestrate the knowledge that already exists, transforming disconnected observations into coherent narratives, narratives into strategic options, and options into confident decisions.</p><p>Understanding cannot simply be downloaded.<br>It has to be designed.</p><h3>The Executive Meeting Reimagined</h3><p>Three years from now, a CEO asks a deceptively simple question: should we acquire a fast-growing AI company before one of our competitors does?</p><p>The room goes quiet. Product sees a capability gap the acquisition could close. Finance sees an aggressive valuation. Legal sees regulatory exposure. Technology sees integration risk. Sales sees immediate customer demand. Strategy sees a possible shift in the market. Everyone is right. No one is aligned, because everyone is looking at a different piece of reality.</p><p>Before the meeting begins, leadership already has access to something fundamentally different. Not another dashboard, market report, or competitive briefing, but a continuously evolving picture of the environment surrounding the enterprise. Customer expectations, competitive movements, emerging technologies, regulatory developments, hiring trends, capital flows, academic research, partner ecosystems, macroeconomic conditions, and geopolitical events are not presented as independent updates. They are presented as relationships.</p><p>Signals begin reinforcing one another. Patterns become visible. Weak signals emerge long before they become obvious trends. The discussion changes. Executives spend less time debating what happened and more time deciding what should happen next.</p><p>The purpose of organizational intelligence is not to eliminate uncertainty. That will never be possible. Its purpose is to increase confidence: confidence that leadership is responding to the right signals, that investments remain aligned with the direction of the market, that the business is adapting before change becomes disruption, and that decisions are informed by a shared understanding rather than isolated expertise.</p><p>This is the capability companies like Netflix and OpenAI appear to be building. Not because they possess better information than everyone else, but because they recognize that information has stopped being the scarce resource. Shared understanding has become the scarce resource. It deserves a name.</p><h3><strong>Enter Design Intelligence</strong></h3><p>Every discipline is shaped by the problems it was created to solve. Industrial Design emerged because organizations needed to create better products. Interaction Design emerged because people needed more intuitive ways to engage with digital technology. Service Design emerged because organizations realized that products alone were no longer enough. Each discipline expanded the designer’s field of view and asked us to think more broadly than the generation before it.</p><p>Design Intelligence represents another expansion. It does not replace strategy, research, product management, business intelligence, competitive intelligence, or artificial intelligence. It operates across them, creating the connective tissue through which an organization perceives change, constructs meaning, and turns complexity into action.</p><p>Viewed through this lens, the role of the designer changes once again. The designer is no longer simply creating products or orchestrating services. The designer becomes an architect of organizational understanding, shaping how intelligence flows through an enterprise, how meaning is constructed across disciplines, and how better decisions emerge from increasing complexity.</p><p>This is not an abandonment of Service Design. It is the logical continuation of where the profession has been heading. Every major evolution in design has occurred when the object of design expanded. We moved from products to experiences, from experiences to services, and from services to systems. The next step is consistent with that progression.</p><p>From designing systems. <br>To designing intelligence.</p><p>Just as Service Design industrialized the design of services, Design Intelligence seeks to industrialize the design of organizational understanding.</p><p>Not because intelligence replaces systems, but because intelligence is what allows systems to sense, learn, adapt, and create value in a world where change has become the only constant. If Design Intelligence is real, then the implications extend far beyond a new discipline. They force us to reconsider what design itself is becoming.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*yDIlsC5edKvVQH-mystjVQ.png" /><figcaption>Some of the world’s most admired products are the visible outcome of invisible organizational understanding.</figcaption></figure><h3><strong>The Next Frontier</strong></h3><p>Every profession eventually reaches a moment when it must decide whether to defend its past or embrace its future. Architecture moved beyond buildings to shape cities. Software engineering evolved from writing code to designing platforms. Strategy expanded from annual planning to continuous adaptation. Design has experienced similar transitions, each challenging the boundaries of what the profession believed itself to be and ultimately making it more valuable.</p><p>I believe we are approaching another of those moments. Not because Service Design has reached its limits, but because the capability it has been developing for decades is becoming essential to every organization navigating an increasingly uncertain world.</p><p>What organizations increasingly need is the ability to make sense of complexity before complexity makes decisions for them. They need to recognize weak signals before they become disruptive forces, connect information before it becomes overwhelming, create shared understanding before strategy is written, and build confidence before capital is allocated.</p><p>The organizations that define the next decade are unlikely to be those with the best products alone, the best services, or even the most sophisticated artificial intelligence. They will be the organizations capable of continuously sensing change, creating shared understanding, making confident decisions, and adapting faster than the environments surrounding them.</p><p>The organizations that think will win.</p><p>That is why the signals from Netflix, OpenAI, and now the U.S. Intelligence Community matter. Not because these organizations are building the same function. They aren’t. They matter because very different institutions are converging on a similar need: the ability to connect complexity, create shared understanding, and translate that understanding into action.</p><p>Here’s the thing about Apple. <br>People often describe Apple’s success through the lens of extraordinary products or through the creative genius of a few. I’ve increasingly come to think that’s only half the story.</p><p>They’re better because the organization has spent decades building extraordinary internal understanding between hardware, software, supply chain, design, operations, developers, retail, ecosystem, and customer behavior.</p><p>Products can be copied. Services can be replicated. Systems can be engineered. Understanding must be cultivated.</p><p><strong>The future will not belong to the organizations with the most intelligence. It will belong to the organizations that are most intelligent.</strong></p><p><em>(This article draws on publicly available sources and industry research. No proprietary or confidential information has been used. AI tools were used selectively to support research, editing, and visual development; the arguments and perspectives expressed are my own.)</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e1cd9fea045c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Service Design Pays. The Advantage Is About to Compound.]]></title>
            <link>https://medium.com/@GregLakloufi/service-design-pays-the-advantage-is-about-to-compound-db1bd7e1ed12?source=rss-2fa4496fe40b------2</link>
            <guid isPermaLink="false">https://medium.com/p/db1bd7e1ed12</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[business-strategy]]></category>
            <category><![CDATA[digital-transformation]]></category>
            <category><![CDATA[service-design]]></category>
            <category><![CDATA[customer-experience]]></category>
            <dc:creator><![CDATA[Greg Lakloufi]]></dc:creator>
            <pubDate>Tue, 06 Oct 2026 16:08:57 GMT</pubDate>
            <atom:updated>2026-10-06T16:12:10.387Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*XyywqtbuauX77nO8nqY8FA.png" /></figure><h4>Organizations that design services end to end will use AI to scale coherence. The rest will scale fragmentation.</h4><p>A customer calls because an order has gone wrong. An employee searches three systems to find the answer. Another team processes a refund. Operations investigates the exception. Finance records the cost.</p><p>Every function handled its part of the work. Every team may even have met its target. The customer still had a terrible Tuesday.<br>Congratulations, everyone.</p><p><strong>The service still fails.</strong></p><p>That is the paradox at the center of many organizations. Customers experience one company, but the company manages their experience through separate functions, systems, budgets and measures.</p><p>The consequences appear everywhere and belong nowhere.</p><p>That is also where I would begin the business case for Service Design: with the cost of what happens today and the opportunity to make it work better.</p><p>Service Design pays when it changes the economics and performance of an experience. It helps an organization see where customers get stuck, where employees compensate, where money leaks from the system and where a better service could create growth.</p><p>The return does not come from the journey maps, blueprints or workshops.<br>Those are business instruments.</p><p>The return comes from what the organization changes because it can finally see and redesign the service as a whole.</p><p>An organization that does this once can create measurable value. An organization that learns to do it repeatedly builds something more consequential: <strong>a competitive capability.</strong> <br>Each redesigned touchpoint strengthens its ability to see, decide and change the next one. That is where the advantage begins to compound.</p><h3><strong>The Chart that put Design on the Executive Agenda</strong></h3><p>For years, the design profession has pointed to the Design Management Institute’s Design Value Index as evidence that design matters commercially. Its 2015 edition reported that <strong>a selected group of 16 design-led companies outperformed the S&amp;P 500 by 211% over a ten-year period.</strong> <strong>[1]</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DnGW_WuXeXML-9m5OJSviQ.png" /></figure><p>McKinsey widened the evidence base in 2018. Its study tracked 300 publicly listed companies across medical technology, consumer goods and retail banking for five years. <strong>Companies in the top quartile of the McKinsey Design Index achieved revenue growth 32 percentage points higher, and total shareholder-return growth 56 percentage points higher, than their industry counterparts. [2]</strong></p><p>Those studies changed the conversation. <strong>Design could finally be discussed as a business capability rather than a mere aesthetic contribution.</strong></p><p>They also left Service Design with a difficult question.</p><p>The research examined <strong>mature design practices</strong> at the company level. It showed an association with stronger business performance. <strong>It could not isolate what Service Design contributed or show where the return appeared inside the enterprise.</strong></p><p>So perhaps we have been asking the question at the wrong level.</p><h3><strong>Nobody asks for the ROI of Marketing in General</strong></h3><p>Apparently, Service Design must arrive carrying a calculator, a laminated business case and documentary evidence of every dollar it intends to produce. Marketing rarely receives quite the same welcome. Why the double standard?</p><p>My friend Marc Stickdorn once framed the question in a way that has stayed with me:</p><blockquote><strong>“We rarely ask for the ROI of marketing as an entire discipline. We evaluate the performance of specific campaigns, investments and decisions.</strong></blockquote><blockquote><strong>Why should Service Design be treated differently?”</strong></blockquote><p>His point is that Service Design should be measured using the same logic: identify the intervention, understand what the organization invested and measure what that intervention produced. [3]</p><p>Companies fund a marketing campaign with a defined audience, objective, cost and set of measures. They evaluate acquisition, conversion, revenue or another intended outcome.<br>They do not wait for one universal number proving that marketing, as an entire discipline, pays.</p><p>The same should be true of Service Design.</p><p>The measurable unit is the intervention: this onboarding journey, this claims process, this support service, this employee experience or this new service proposition. The question becomes:</p><p><em>What changed after we redesigned this service?</em></p><p>That is a question leaders can evaluate.</p><p>It also moves Service Design beyond the exhausting task of defending an entire discipline. The question is no longer whether Service Design pays in general. It is where to apply it next, which interventions create the greatest value and how quickly the organization can turn what it learns into an advantage elsewhere.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tXOezqPitk4uZ0biISlmLw.png" /></figure><h3><strong>All the Dashboards are Green. The Service is Still Broken.</strong></h3><p>There is another reason the value of Service Design is hard to see.</p><p><strong>The cost of a failed experience rarely appears in one place.</strong></p><p>Marketing pays to acquire the customer. Sales records the transaction. Digital measures completion and deflection. Operations manages exceptions. The contact center tracks handling time. Finance sees refunds and credits. Technology funds patches and new platforms. Human Resources sees the workload and turnover created when employees compensate for broken processes.</p><blockquote>Each function can improve its own metric while the experience performs worse end to end.</blockquote><p>A digital team can increase self-service adoption while more customers fail to complete the task. A contact center can reduce handling time while repeat calls increase. Operations can meet its service-level agreement while employees rely on manual workarounds. Marketing can improve acquisition while onboarding failure accelerates churn.</p><p><strong>This is one of the small miracles of modern management: every dashboard can be green while the customer is still seeing red.</strong></p><p>This is the hidden economics of service failure. <br>Value is created and lost <strong>across</strong> organizational boundaries, while most companies measure performance <strong>within</strong> them.</p><p>Together, these losses form <strong>the hidden cost of fragmentation</strong>. It appears as abandonment, failure demand, rework, employee burden, avoidable technology spending, customer loss and change that never quite delivers its promised value.</p><p>No finance system contains a single line item for it. <br>That would be far too convenient. The cost arrives disguised as repeat calls, duplicated work, avoidable escalations, employee exhaustion and customers quietly deciding they have had enough.</p><p>No single function sees the whole cost.<br><strong>The enterprise pays all of it.</strong></p><h3><strong>What if we Measured the Service Instead?</strong></h3><p>Service Design makes those distributed consequences visible as one connected system.</p><p>It links customer behavior, employee effort, operational performance, technology and financial outcomes. Its distinctive contribution is larger than producing another set of customer insights. It shows how the parts combine to create value or destroy it.</p><p>Consider a contact-center problem. <br>Customers keep calling to ask where their orders are. The immediate response might be to improve agent training, rewrite the help page, reduce average handling time or introduce a chatbot.</p><p>Each could help.</p><p>But suppose the calls originate in unreliable delivery estimates, conflicting notifications and systems that cannot share an accurate order status. The contact center is absorbing a problem created elsewhere.</p><p><strong>Optimizing the call treats the visible symptom.<br>Designing the service examines the sequence that produced it.</strong></p><p>It connects what customers experience in the frontstage with the work happening backstage: information flows, policies, responsibilities, systems, incentives and handoffs.</p><p>The commercial question changes:</p><p><em>Which contacts can we prevent by making the underlying service work, and which contacts should we make easier to resolve?</em></p><p><strong>A faster answer is valuable when the customer’s need is resolved.<br>A call that never needs to happen may be more valuable still.</strong></p><h3><strong>Make the Value Chain Visible</strong></h3><p>Claims about Service Design become credible only when we can trace how an intervention produces an outcome.</p><p>The chain should be visible:</p><h4><strong>• Design Intervention:</strong></h4><p>Research, service redesign, prototyping or implementation changes something specific.</p><h4><strong>• Service Behavior:</strong></h4><p>A handoff disappears. Information arrives earlier. A policy changes. Customers or employees can complete a task differently. A system supports the work more effectively.</p><h4><strong>• Human and Operational Outcome:</strong></h4><p>Abandonment falls. Resolution improves. Employees perform less rework. Processing becomes faster. Errors or complaints decline.</p><h4><strong>• Business Measure:</strong></h4><p>Conversion, retention, cost to serve, productivity, adoption or implementation performance changes.</p><h4><strong>• Economic Value:</strong></h4><p>The organization creates revenue, prevents loss, removes cost, releases capacity or avoids an unnecessary investment.</p><p>If we cannot trace that chain, an ROI claim is advocacy.</p><p><strong>If we can trace it, Service Design becomes an investment proposition that can be measured, tested and improved.</strong></p><h3><strong>Six Pathways to Value</strong></h3><p>The measures will change with the experience, but six core value pathways recur:</p><h4><strong>1. Revenue and Conversion</strong></h4><p>Customers abandon applications, purchases and activations when an experience creates uncertainty or fails between channels. Service Design traces those barriers across the journey and the operating system behind it. Relevant measures may include completed transactions, activation, conversion and incremental contribution margin.</p><h4><strong>2. Retention and Loyalty</strong></h4><p>The promise made during marketing and sales can break down during onboarding, delivery, billing or support. Redesigning the relationship across those moments can influence churn, renewal, repeat purchase and customer lifetime value.</p><h4><strong>3. Cost to Serve</strong></h4><p>Repeated contacts, avoidable escalations, manual workarounds and rework are often symptoms of am experience that does not function end to end. Service Design can reveal the source of that demand and move the solution upstream. Measures may include cost per successfully resolved need, repeat-contact rates, processing time, errors and demand avoided.</p><p>Time released becomes capacity. It becomes a cash saving when expenditure actually falls.</p><h4><strong>4. Investment Risk</strong></h4><p>Organizations regularly fund technology or transformation programs before confirming that the proposed solution addresses the underlying need. Research, prototyping and testing can change the decision before the largest costs are committed. The value may appear as avoided expenditure, a stronger investment case, a redesigned scope or an early decision to stop.</p><h4><strong>5. Workforce Performance</strong></h4><p>Employees often hold broken services together. They search across systems, interpret conflicting policies and invent workarounds so customers receive something close to the intended experience. Redesigning the experience can reduce task time, errors, rework and workload while improving the employee’s ability to help the customer.</p><h4><strong>6. Implementation Performance</strong></h4><p>A promising service concept creates little value if the organization cannot adopt and operate it. Service Design involves the people who deliver the experience while it is being designed and tested. Adoption, time to proficiency, implementation speed and live-service performance show whether the change survived contact with the organization.</p><p>These pathways do not promise the same return in every situation.<br>They tell us where to look.</p><p>Each successful intervention can also leave behind more than its immediate result. It builds reusable knowledge, stronger cross-functional relationships, better measures and greater confidence in the organization’s ability to change. The next experience can be redesigned faster and with less uncertainty.</p><p><strong>Competitors beginning from scratch do not merely lack a better service. They lack the organizational muscle that produced it.</strong></p><h3><strong>The Evidence is Imperfect. It is Also Real.</strong></h3><p>One peer-reviewed case examined a Service Design project at Zodiac Training, a UK vocational training provider. The organization redesigned and implemented a digitally mediated training service that included learner sign-up, evidence submission, progress tracking, quality monitoring and claims. [4]</p><p><strong>Five months after implementation, 41% of learners were using the new service and the organization estimated annual net cost savings of £125,000.</strong> When the researchers returned the following year, more than 4,000 learners were using it and on-time completion had improved. [4]</p><p>The authors were part of the design and evaluation team, and they explicitly acknowledged the limits of a single retrospective case. That restraint matters. The case does not prove a universal return from Service Design. <strong>It demonstrates how redesigning a specific experience can create measurable operational, financial and stakeholder value.</strong></p><p>A more recent Bosch Service Solutions case describes a telecommunications engagement involving fragmented systems, manual tasks and problems with first-call resolution, handling time, customer satisfaction and cost efficiency. The team mapped front-office and back-office processes, identified opportunities for self-service and automation, and applied bots and decision trees across several customer processes. [5]</p><p><strong>Bosch reports that average handling time fell by three minutes per case, described in its results as a 20% reduction, while overall customer satisfaction improved. [5]</strong></p><p>This is a provider-reported case involving an unnamed client, so its claims should be read accordingly. But the value chain is visible: understand the experience, redesign the work, implement the change and measure the result.</p><p>The Design Council’s Design Value Framework reinforces the need to look beyond one financial measure. It identifies financial, social, environmental and democratic value, and suggests indicators such as savings from prototyping, improved investment readiness and earlier identification of scalable opportunities. [6]</p><p>The evidence base remains more fragmented than Service Design advocates might wish.</p><p>That is an argument for better measurement.<br>It is not an argument for ignoring the cost of services that are already failing.</p><h3><strong>What Happens if Nobody Designs the Service?</strong></h3><p>Of course, an organization can choose not to intentionally design its services/experiences.</p><p>But the experiences will <strong>STILL</strong> be designed.</p><p>Customers and employees will do it themselves, one workaround, escalation and improvised spreadsheet at a time.</p><p>This is the organization designing around itself.</p><p>It is designed by the accumulation of other decisions: the org chart, performance targets, procurement rules, legacy technology, channel ownership, compliance interpretations and whichever transformation program arrived most recently.</p><p><strong>No one intends to create a fragmented experience. Each function makes a rational decision within its own boundaries.</strong></p><p><strong>But the customer experiences the accumulated result as one company.</strong></p><p>Over time, the consequences compound. Customers abandon journeys or leave without explaining why. Failure demand enters the contact center. Employees absorb complexity through workarounds. Technology teams build around symptoms. New platforms preserve old fragmentation at greater scale. Leaders fund another transformation because the previous one changed the interface without changing the service.</p><p>The organization becomes highly efficient at compensating for problems it never solves. That is an operating cost.</p><p><strong>Eventually, it becomes a strategic risk.</strong></p><p>A company that cannot see how its experience works end to end struggles to change it. It moves slowly when customer expectations shift, when a competitor removes friction or when technology enables a different service model.</p><p>Its legacy is no longer only technical. It is embedded in policies, incentives, workflows and decisions distributed across the enterprise.</p><p>Former Telefónica Deutschland CEO Thorsten Dirks once put it more bluntly. Translated from the German:</p><blockquote>“If you digitize a shitty process, you get a shitty digital process.” [7]</blockquote><p>We have seen this movie before. <br>We just gave the sequel a much larger technology budget.</p><h3>AI Will Scale the Experience You Already Have. Are You Sure You Want More of it?</h3><p>Organizations are preparing to automate experiences and services they have never fully designed end to end.</p><p>If the experience is coherent, AI can make it faster, more responsive and more personal. But if it is currently fragmented, AI can automate the workaround, accelerate the wrong decision and make the underlying failure harder to see.</p><p>A chatbot may reduce contact volume while concealing unresolved demand. An agent assistant may shorten calls while the failure that generates them continues. Automation may simply move work from employees to customers and call the result efficiency.</p><p><strong>Digitization preserved broken processes. <br>AI can industrialize broken services.</strong></p><p>This is where the competitive gap begins to widen. Organizations with coherent experiences can use AI to scale what works.</p><blockquote><strong>Organizations without coherent experiences today may spend the next decade becoming extraordinarily efficient at doing the wrong things.</strong></blockquote><p>Before asking where AI can be inserted, leaders need to understand why the demand exists, what outcome the customer is trying to achieve, what employees must do to deliver it and which parts of the operating system create the failure.</p><p><strong>Service Design provides that view. <br>It gives organizations the chance to improve the experience before they scale it.</strong></p><h3>What Compounding Advantage Looks Like</h3><p>Imagine an enterprise that has learned to do this repeatedly. <br>It treats its major experiences and services as operating assets.</p><p>It gives each service clear end-to-end accountability.</p><p>It measures successful customer and employee outcomes alongside functional efficiency.</p><p>It establishes the value chain before transformation begins.</p><p>It connects frontstage experience with backstage operations, policy, data and technology.</p><p>It tests the service before committing the largest investments.</p><p>It uses AI to remove failure demand and expand human capability rather than to conceal fragmentation.</p><p>In that organization, Service Design is not a specialist activity waiting to be invited into a project. It is part of how the enterprise decides what to change, what to fund and how to know whether the change worked.</p><h3><strong>Start With One Experience</strong></h3><p>Leaders do not need to settle the value of an entire profession before making an investment decision.</p><p>Start with one experience/service where the stakes are visible: an expensive onboarding journey, recurring complaints, high repeat-contact volumes, poor conversion, employee workarounds or a transformation customers struggle to use.</p><p>Establish the baseline. Identify what is failing, who experiences it, what it costs and which outcomes matter. Select measures before redesign begins. Test a change, then compare performance with the baseline and, where feasible, a comparable group.</p><p>Use commercial and service measures together.</p><p>Lower handling time matters when customers get their needs resolved.</p><p>Higher digital adoption matters when customers can successfully complete the task.</p><p>Improved satisfaction matters, but it should not automatically be converted into revenue.</p><p>Separate measured results from estimates. Distinguish released capacity from realized savings and planned expenditure avoided from cash already saved. Give each expected benefit an owner who can see it through implementation.</p><p>The result is a disciplined investment proposition: a defined service problem, a credible value pathway, a bounded intervention and evidence of what changed.</p><h3>The Cost is Already There</h3><p>Service Design is often asked to prove its value while the cost of fragmented services goes largely unmeasured.</p><p>Look at where customers get stuck.</p><p>Look at where employees compensate.</p><p>Look at where the organization pays repeatedly to correct problems that originate somewhere else.</p><p>Those are signals that the service is producing consequences nobody owns end to end.</p><blockquote>The question is no longer only whether your organization can afford to invest in Service Design. It is whether it can afford to keep leaking value while competitors learn to design, measure and improve their services as whole systems.</blockquote><p>AI will widen that gap. It will scale coherence where coherence exists and will increase fragmentation where it does not.</p><p>The customer already experiences one company.</p><p>The company should learn to design like one.</p><h4><strong>Sources</strong></h4><p>[1] Rae, J. Design Value Index Exemplars Outperform the S&amp;P 500 Index (Again) and a New Crop of Design Leaders Emerge. Design Management Review, 27, 4–11. <a href="https://doi.org/10.1111/drev.12040">https://doi.org/10.1111/drev.12040</a></p><p>[2] Sheppard, B., Sarrazin, H., Kouyoumjian, G. and Dore, F. The business value of design. McKinsey &amp; Company, October 25, 2018. <a href="https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/the-business-value-of-design">https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/the-business-value-of-design</a></p><p>[3] Stickdorn, M. Ask Marc: measuring service design impact. Smaply, February 16, 2021. <a href="https://www.smaply.com/blog/ask-marc-measuring-service-design">https://www.smaply.com/blog/ask-marc-measuring-service-design</a></p><p>[4] Lievesley, M. and Yee, J. Valuing Service Design: Lessons from SROI. DRS International Conference 2012. <a href="https://doi.org/10.21606/drs.2012.74">https://doi.org/10.21606/drs.2012.74</a></p><p>[5] Bosch Service Solutions. Case study: Higher customer satisfaction thanks to a significant increase in speed to answer. Accessed October 5, 2026. <a href="https://www.boschservicesolutions.com/en/in-the-spotlight/telecommunication/">https://www.boschservicesolutions.com/en/in-the-spotlight/telecommunication/</a></p><p>[6] Design Council. The Design Value Framework. Accessed October 5, 2026. <a href="https://www.designcouncil.org.uk/resources/the-design-value-framework/">https://www.designcouncil.org.uk/resources/the-design-value-framework/</a></p><p>[7] Bitkom. <em>To Get to Transformational Products, IT Needs Design.</em> Quotation by Thorsten Dirks. <a href="https://www.bitkom.org/sites/main/files/2018-11/7_Gassner_To%20get%20to%20transformational%20products%2C%20IT%20needs%20Design.pdf?utm_source=chatgpt.com">https://www.bitkom.org/sites/main/files/2018-11/7_Gassner_To%20get%20to%20transformational%20products%2C%20IT%20needs%20Design.pdf</a></p><p><em>(This article draws on publicly available sources and industry research. No proprietary or confidential information has been used. AI tools were used selectively to support research, editing, and visual development; the arguments and perspectives expressed are my own.)</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=db1bd7e1ed12" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Artificial Intelligence Needs Systems Intelligence]]></title>
            <link>https://medium.com/@GregLakloufi/artificial-intelligence-needs-systems-intelligence-5128d573bc2c?source=rss-2fa4496fe40b------2</link>
            <guid isPermaLink="false">https://medium.com/p/5128d573bc2c</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[ai-governance]]></category>
            <category><![CDATA[organizational-design]]></category>
            <category><![CDATA[systems-thinking]]></category>
            <category><![CDATA[strategic-design]]></category>
            <dc:creator><![CDATA[Greg Lakloufi]]></dc:creator>
            <pubDate>Mon, 14 Sep 2026 17:14:37 GMT</pubDate>
            <atom:updated>2026-09-16T19:35:49.832Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*HT9RJWfpBDIrY5qe_bQXZA.png" /></figure><h4>Why safer models will not be enough, and why the institutions surrounding them must learn at machine speed.</h4><p>The most important warning about artificial intelligence may not be that machines are becoming too intelligent.</p><p>It may be that the organizations surrounding them are not.</p><p>AI capabilities are advancing at extraordinary speed. Models are becoming more autonomous, more connected, and more capable of acting across the digital world. They are increasingly contributing to the development of the systems that will succeed them, compressing the distance between one generation of capability and the next.</p><p>But the institutions responsible for governing those capabilities still operate through annual strategies, periodic audits, fragmented oversight, fixed policies, and decisions distributed across organizational boundaries.</p><p>The capability is advancing continuously.<br>The organization is not.</p><p>That difference may become one of the defining strategic challenges of the next decade.</p><p>It is tempting to call this an AI safety problem. In part, it is. It is also a policy problem, a cybersecurity problem, an economic problem, and increasingly a national security problem.</p><p>But underneath all of those is something larger.|<br>It is a systems problem.</p><p>And systems problems cannot be solved by improving one component at a time.</p><h3>The Frontier is Beginning to Accelerate Itself</h3><p>In his recent essay, <a href="https://darioamodei.com/post/we-must-pace-the-frontier"><em>We Must Pace the Frontier</em></a>, Anthropic CEO <a href="https://www.linkedin.com/in/dario-amodei-3934934/">Dario Amodei</a> argues that frontier AI development may need to become more deliberate. His concern is not simply that AI is becoming more capable. It is that capability may be accelerating beyond our ability to understand, evaluate, and control it.</p><p>Amodei describes this as the early formation of recursive self-improvement: AI systems increasingly contributing to the development of their successors. The technology is not autonomously redesigning itself in some science-fiction loop, at least not yet. But AI is already helping researchers write code, analyze results, and accelerate the work required to build the next generation.</p><p>This changes the tempo of development.</p><p>Capability is no longer moving only at the familiar speed of research cycles, product releases, or institutional adoption. The development cycle itself is beginning to accelerate. Risk and consequence may now evolve more quickly than the organizations charged with interpreting them.</p><p>In an earlier essay, Amodei described humanity as being handed almost unimaginable power, then questioned whether our “social, political, and technological systems possess the maturity to wield it.”</p><p>That is an extraordinary statement from one of the people closest to the frontier.</p><p><strong>It is also more than a warning about the technology.<br>It is an indictment of the system around it.</strong></p><h3>The Model is Not the System</h3><p>Much of the current AI conversation focuses on the model.</p><p>Is it accurate? Is it aligned? Is it secure? Can it reason? Can it deceive? Can it escape a controlled environment? Can we understand what is happening inside it?</p><p>These are critical questions. But the model is not the whole system.</p><p>The system also includes the people who build it, the data used to train it, the environments in which it is trained and tested, the vendors supporting those environments, and the incentives shaping development and release decisions.</p><p>It includes the evaluations used to judge it, the infrastructure through which it operates, the security protecting it, the organizations deploying it, the regulators overseeing it, and the people affected when it acts.</p><p>It includes what each participant can see.</p><p>It includes what they cannot see.</p><p>It includes who has the authority to intervene, what evidence is required, how quickly a decision can be made, and what happens when commercial pressure, public safety, and national security point in different directions.</p><p>The behavior of AI will not emerge from the model alone. It will emerge from the interaction between the model and this surrounding system.</p><p>That distinction matters because even technically sophisticated systems can fail for remarkably ordinary reasons.</p><p>Amodei makes this point directly: <a href="https://darioamodei.com/post/we-must-pace-the-frontier">“Many things go wrong not because companies are missing some important theory or insight, but because of problems in execution.”</a></p><p>Broken training environments. Imperfect data filtering. Weak monitoring. Inadequate sandboxing. Vendor failures. Security gaps. Misaligned incentives. Ambiguous authority. Information that never reaches the people empowered to act.</p><p>These are not isolated technical defects.<br>They are symptoms of a system that has not been designed to understand itself.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DWiASmIho99V0uGDxL9ctg.png" /></figure><h3>The Governance Gap is a Design Gap</h3><p>Artificial intelligence is advancing through continuous experimentation.</p><p>Most institutions still operate through episodic governance.</p><p>AI teams run evaluations, observe behavior, adjust models, change training environments, and deploy capabilities in rapid succession. The surrounding organization often responds through scheduled reviews, static control frameworks, annual risk assessments, and committees separated from the work by layers of reporting.</p><p>One system learns by the hour.<br>The other learns by the quarter.<br>This is not simply a difference in pace. It is a structural mismatch.</p><p>When technology changes faster than the institution can interpret the consequences, governance becomes retrospective. Policies describe yesterday’s system. Audits examine yesterday’s evidence. Controls address yesterday’s failure modes.</p><p>The organization may appear governed while becoming progressively less capable of governing what the technology has become.</p><p>Amodei proposes serious interventions: permanent external evaluators with employee-like access, capability-based checkpoints, stronger operational security, improved interpretability, more rigorous testing, industry standards, government coordination, and eventually international agreements.</p><p>These mechanisms move the conversation beyond principles and toward implementation.</p><p>But a list of safeguards is not yet a system.</p><p>Who determines which signals matter? Who distinguishes an anomaly from an emerging pattern? How does an evaluator’s observation travel across technical, legal, operational, and executive boundaries?</p><p>What threshold converts concern into action? Who has the authority to slow, stop, or redirect development? How are lessons shared across organizations without exposing intellectual property or creating new security vulnerabilities?</p><p>How does a regulator evaluate a capability that may have changed materially between the beginning and end of the review?</p><p>These are not administrative details to address after the technical work is complete. They determine whether the safeguards work at all.</p><p>Governance is often treated as a layer placed over technology. A policy is written. A board is established. A review is required. A checklist is completed.</p><p>But governance is not a layer.</p><p>It is a living system of sensing, interpretation, decision, intervention, and learning.</p><p>AI developers, policymakers, security teams, executives, and designers are each building necessary parts of the response. What remains underdeveloped is the connective capability that allows those efforts to operate as one adaptive institutional system.</p><p>Closing that gap requires more than another control. It requires a practice capable of redesigning how the institution itself behaves.</p><p><strong>The governance gap is a design gap.</strong></p><h3>Where Systems Intelligence Meets Strategic Design</h3><p>Systems Intelligence did not begin with artificial intelligence.</p><p><a href="https://www.linkedin.com/in/raimohamalainen/">Raimo Hämäläinen</a> and <a href="https://www.linkedin.com/in/esa-saarinen-0bb7a11b/">Esa Saarinen</a> introduced the concept more than two decades ago to describe intelligent human behavior within complex systems. Their work explored how people can perceive their participation in a larger whole, understand the feedback surrounding them, and act productively even when the system cannot be completely known or controlled.</p><p>That foundation matters.</p><p>Systems Intelligence has since been applied to organizational development, <a href="https://onlinelibrary.wiley.com/doi/full/10.1002/hrdq.21455?utm_source=chatgpt.com">learning organizations</a>, engineering education, and design. In 2021, Hämäläinen and Saarinen joined <a href="https://www.linkedin.com/in/j-tuomas-harviainen-27068739/">J. Tuomas Harviainen</a> in <a href="https://research.aalto.fi/en/publications/system-intelligence-as-an-organizational-approach-for-service-des/?utm_source=chatgpt.com">bringing its principles directly into service design</a>, emphasizing a pragmatic, people-centered approach to improving organizational systems. This existing body of work should be recognized, not displaced.</p><p>But the convergence of artificial intelligence, institutional complexity, and strategic risk now presents a larger design challenge.</p><p>The question is no longer only how people can act intelligently within complex systems.</p><p>It is also how we intentionally design the structures, technologies, relationships, governance, and learning mechanisms through which an organization can behave more intelligently as a whole. That second question is the extension I am proposing here.</p><p>This is where Systems Intelligence meets Strategic Design.</p><p>Systems Intelligence helps people perceive and act intelligently within complexity. Strategic Design helps us deliberately reshape the conditions producing that complexity.</p><p>Together, they form an applied practice for designing more intelligent and adaptive institutions.</p><p><strong>Strategic Design &amp; Systems Intelligence, or SDSI, is the applied practice of perceiving, designing, and deliberately shaping complex human, organizational, and technological systems so that institutions can sense, interpret, decide, act, and learn as coherent adaptive wholes.</strong></p><p>SDSI brings together systems thinking, strategic design, organizational design, systems engineering, technology, business strategy, and futures thinking.</p><p>It does not attempt to make complexity controllable.<br>It makes complexity intelligible enough to act upon.</p><p>This may sound ambitious, but in practice, its logic is already visible in some of the world’s most effective operating systems.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gD9RNmWH9RZp81-4Qi7Wvg.png" /></figure><h3>The Power was Never in the Cord</h3><p>Long before artificial intelligence, Toyota designed one of the clearest examples of an organization capable of seeing and responding to its own behavior.</p><p>In the Toyota Production System, the principle of <em>jidoka</em> allows equipment to stop when an abnormality occurs. An andon display makes the condition visible so people can respond, investigate the cause, and prevent the defect from continuing through production.</p><p>At first glance, it appears to be a signaling mechanism.</p><p>A light turns on. A machine stops. Someone asks for help.</p><p>But the signal is only the visible artifact.</p><p>The real innovation is the institutional architecture surrounding it.</p><p>The system determines what qualifies as an abnormality. It makes the problem visible where the work is happening. It creates the authority to interrupt production. It mobilizes people capable of interpreting the problem. It supports intervention before the defect propagates. It turns what happened into knowledge that can improve the larger production system.</p><p>The power of the andon system was never the signal.</p><p>It was the organization’s capacity to respond intelligently to what the signal revealed.</p><p>This is what many AI governance structures currently lack. Organizations are adding policies, dashboards, review boards, principles, and controls. But the connection between signal, understanding, authority, and intervention often remains fragmented.</p><p>A model produces an unexpected result. A frontline employee notices an unusual pattern. An evaluator discovers a capability the development team did not anticipate. A customer is harmed by a decision no one function fully owns.</p><p>The signal exists.</p><p>The surrounding institution does not know how to respond.</p><p>What is the andon cord for artificial intelligence?</p><p>Who can pull it? What becomes visible when they do? Who arrives to interpret the problem? What authority do they possess? What happens to the evidence afterward?</p><p>Those questions reveal whether an organization has established AI governance or merely accumulated governance artifacts. Together, those questions resolve into five connected institutional capabilities.</p><h3>The SDSI Operating Model</h3><h4>1. Sense</h4><p><strong>The institution must see what is actually happening across the system, not merely what each component reports.</strong></p><p>That requires continuous technical evaluation, operational monitoring, independent observation, frontline reporting, external signals, and visibility into the environments where models are trained and deployed.</p><p>Amodei’s proposal for embedded external evaluators is powerful for this reason. It adds an independent sensing function inside a system otherwise dependent on self-reporting.</p><p>But more information does not automatically produce more intelligence.</p><p>The organization must decide what it needs to observe, where meaningful signals are likely to emerge, and how those signals reach the right people before they become incidents.</p><h4>2. Interpret</h4><p><strong>Data provides signals about what is happening. Intelligence develops an understanding of why those signals matter.</strong></p><p>An unexpected model behavior may be a technical anomaly, an incentive problem, a security vulnerability, an emerging capability, or evidence that an evaluation no longer measures what its designers believe it measures.</p><p>No single discipline can interpret all of those possibilities.</p><p>Technical expertise must connect with operational experience, behavioral insight, business context, ethics, policy, security, and geopolitical understanding.</p><p>The challenge is not collecting more perspectives.</p><p>It is integrating them into shared understanding.</p><h4>3. Decide</h4><p><strong>Intelligence without decision authority creates awareness without consequence.</strong></p><p>Organizations must define who can act, under what conditions, and using what evidence. Those decision rights must evolve with the capability rather than remain frozen in an organizational chart.</p><p>This is where capability-based checkpoints become important. Governance should not wait for the next quarterly meeting if a system crosses a meaningful threshold tomorrow.</p><p>The evidence should govern the tempo.</p><h4>4. Intervene</h4><p><strong>Complex systems rarely offer a single control point.</strong></p><p>An intervention may involve changing a model, restricting access, redesigning a workflow, altering an incentive, isolating an environment, changing procurement requirements, slowing deployment, or reshaping the relationship between human and machine authority.</p><p>The objective is not to control every variable. That is impossible.</p><p>The objective is to identify where intervention can meaningfully change system behavior without creating worse second-order consequences elsewhere.</p><h4>5. Learn</h4><p><strong>Every intervention creates new evidence.</strong></p><p>Institutions must capture that evidence, translate it across boundaries, update their assumptions, and redesign the surrounding system accordingly.</p><p>Learning cannot be confined to a retrospective after the work is complete because, with AI, the work may never be complete. The governance model must learn as continuously as the capability changes.</p><p><strong>Sense. Interpret. Decide. Intervene. Learn.</strong></p><p>These capabilities do not replace technical evaluation or established risk-management frameworks. They describe the institutional connective tissue that allows those disciplines to produce coherent action.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*vz5AfkxOZehQXnMR36et7A.png" /></figure><h3>What Aerospace Already Understands</h3><p>Aerospace learned long ago that the performance of a sophisticated aircraft cannot be separated from the larger mission system around it.</p><p>An advanced aircraft may contain extraordinary technology, but mission success also depends on sensors, communications, software, intelligence, logistics, maintenance, command structures, human judgment, allied interoperability, and the ability to adapt as conditions change.</p><p>The platform matters.</p><p>The relationships determine what the platform can accomplish.</p><p>A modern fighter is not simply an aircraft. It is an intelligent node inside a system of systems. The thing that detects a threat may not be the thing that interprets it. The thing that interprets it may not be the thing authorized to act. The thing that acts may depend on another part of the system to measure the result.</p><p>Capability emerges through coordination.</p><p>I explored this transition more deeply in <a href="https://medium.com/@GregLakloufi/the-system-is-the-design-f17ee8bebd41"><em>The System is the Design</em></a>, examining aerospace’s evolution from standalone platforms into adaptive systems of systems and drawing on a conversation with <a href="https://www.linkedin.com/in/togatrew/">Colonel Jason Trew, PhD, U.S. Air Force (Ret.).</a></p><p>The argument was simple: <strong>A plane is no longer merely a plane.</strong> It is a sensing, learning, and decision node whose value depends on everything surrounding it.</p><p>That essay examined the system from the perspective of capability. The question here is what happens when the same logic is applied to governance. If the platform has become a system, the institution responsible for it must become one as well.</p><p>AI is moving in the same direction. The model is becoming one intelligent node inside a much larger system of people, machines, organizations, and decisions.</p><p>This is already visible in defense strategy. <a href="https://www.linkedin.com/in/craighmartell/">Craig Martell</a>, then the Department of Defense’s Chief Digital and AI Officer, argued that the objective was not to identify a few decisive applications, but to strengthen <a href="https://www.war.gov/News/Releases/Release/Article/3578219/dod-releases-ai-adoption-strategy/">“the organizational environment within which our people can continuously deploy data analytics and AI capabilities for enduring decision advantage.”</a></p><p>That language matters.</p><p>The competitive advantage is not any single model. It is the environment that can continually discover, test, deploy, govern, and improve models in pursuit of a mission.</p><p><strong>The thing that learns fastest will not necessarily be the machine.<br>It may be the system surrounding it.</strong></p><h3>Pacing is Not the Same as Stopping</h3><p>The debate about AI is often forced into a false choice.</p><p>Accelerate or regulate.<br>Innovate or constrain.<br>Compete or govern.</p><p>But the real strategic question is not whether progress should be fast or slow. It is whether capability, understanding, and institutional control remain in balance.</p><p>Pacing does not need to mean selecting an arbitrary speed limit. It can mean designing thresholds at which new evidence triggers deeper evaluation, broader participation, different authorities, or more constrained deployment.</p><p>It can mean moving quickly when the system demonstrates justified confidence and deliberately when uncertainty becomes consequential.</p><p>The National Institute of Standards and Technology has argued that AI risk management should be <a href="https://nvlpubs.nist.gov/nistpubs/ai/NIST.AI.100-1.pdf">“continuous, timely, and performed throughout the AI system lifecycle.”</a> The Department of Defense has similarly developed responsible AI guidance intended to evolve as a living toolkit rather than remain a fixed compliance artifact.</p><p>The direction is correct.</p><p>But continuous governance requires more than continuous testing.</p><p>It requires an organization capable of turning new evidence into new behavior.</p><p>That is the difference between having a governance process and becoming a learning system.</p><h3>The Institution Must Become Intelligent</h3><p>We often describe intelligence as a property of a person or a machine.</p><p>But intelligence can also become a property of an organization.</p><p>An intelligent institution can detect meaningful change before it becomes obvious. It can connect information across boundaries, challenge its own assumptions, distribute authority without losing accountability, and act before certainty arrives.</p><p>Most organizations were built to produce consistency, control variance, protect expertise, and scale repeatable processes. Those qualities created enormous value in relatively stable environments.</p><p>The same structures become liabilities when conditions change faster than information can move through the hierarchy.</p><p><strong>Artificial intelligence will expose this weakness.</strong></p><p>Organizations will automate decisions without redesigning accountability. They will accelerate work without improving institutional judgment, multiply signals without clarifying who can act, and connect powerful models to operating environments that cannot see across themselves.</p><p>That is not transformation.</p><p>It is acceleration without adaptation.</p><p>AI introduced into a fragmented enterprise does not automatically make the enterprise intelligent.</p><p><strong>It can create a faster dysfunctional system.</strong></p><h3>Design Must Take Responsibility for The Whole</h3><p>If institutional intelligence must be designed, then design’s responsibility must expand.</p><p>We designed objects.<br>Then interfaces.<br>Then experiences.<br>Then services.</p><p>Now we must design the systems that produce them.</p><p>Artificial intelligence makes that progression unavoidable.</p><p>Designers can no longer focus only on whether an AI experience is useful, usable, or desirable. Strategic leaders can no longer focus only on adoption, productivity, or return on investment. Technologists can no longer focus only on model performance. Policymakers can no longer assume that a rule becomes an operational reality simply because it has been written.</p><p><strong>Someone must take responsibility for the coherence of the whole.</strong></p><p>That does not mean creating one centralized authority that attempts to control everything. Complex systems resist that kind of control.</p><p>It means designing the conditions through which the system can see itself.</p><p>It means designing how evidence moves, who participates in its interpretation, where authority sits, and how that authority changes as risk becomes more consequential.</p><p>It means designing environments where emerging capabilities can be tested without immediately becoming operational consequences.</p><p>It means designing implementation pathways that connect insight to action, strategy to execution, and principles to behavior.</p><p>Most importantly, it means designing the capacity to adapt.</p><p>Because the goal is not to build a perfect governance structure for today’s AI.</p><p>It is to build an institution capable of redesigning itself for whatever AI becomes next.</p><h3>The Real Intelligence Test</h3><p>We are building machines that can identify patterns, generate knowledge, support decisions, and increasingly act in the world.</p><p>The technical achievement is extraordinary.</p><p>But the deeper test will not be whether machines appear increasingly intelligent.</p><p>They already do.</p><p>The test will be whether the systems surrounding them become intelligent enough to wield what they have created.</p><p>Safer models, better regulation, stronger evaluation, security, and oversight will all matter. None will succeed as a disconnected intervention. The institutions responsible for artificial intelligence must be able to sense, interpret, decide, intervene, and learn as one adaptive system.</p><p>Artificial intelligence will not be made safe, useful, or governable by technical intelligence alone.</p><p>It will depend on Systems Intelligence.</p><p><strong>The future will not be shaped only by those who build the most intelligent machines.</strong></p><p><strong>It will be shaped by those who can intelligently design the systems around them.</strong></p><h3>References</h3><p>Hämäläinen, Raimo P., and Esa Saarinen, editors. <a href="https://research.aalto.fi/en/publications/systems-intelligence-discovering-a-hidden-competence-in-human-act/"><em>Systems Intelligence: Discovering a Hidden Competence in Human Action and Organizational Life</em></a>. Helsinki University of Technology, 2004.</p><p>Amodei, Dario. <a href="https://darioamodei.com/essay/the-adolescence-of-technology">“The Adolescence of Technology.”</a> January 2026.</p><p>Amodei, Dario. <a href="https://darioamodei.com/post/we-must-pace-the-frontier">“We Must Pace the Frontier.”</a> September 2026.</p><p>National Institute of Standards and Technology. <a href="https://nvlpubs.nist.gov/nistpubs/ai/NIST.AI.100-1.pdf"><em>Artificial Intelligence Risk Management Framework (AI RMF 1.0).</em></a> January 2023.</p><p>Toyota Motor Corporation. <a href="https://global.toyota/en/company/vision-and-philosophy/production-system/">“Toyota Production System.”</a></p><p>U.S. Department of Defense. <a href="https://www.war.gov/News/Releases/Release/Article/3578219/dod-releases-ai-adoption-strategy/">“DOD Releases AI Adoption Strategy.”</a> November 2023.</p><p>U.S. Department of Defense Chief Digital and Artificial Intelligence Office. <a href="https://www.war.gov/News/Releases/Release/Article/3588743/cdao-releases-responsible-ai-rai-toolkit-for-ensuring-alignment-with-rai-best-p/">“Responsible Artificial Intelligence Toolkit.”</a> November 2023.</p><p>Harviainen, J. Tuomas, Raimo P. Hämäläinen, and Esa Saarinen. <a href="https://research.aalto.fi/en/publications/system-intelligence-as-an-organizational-approach-for-service-des/">“Systems Intelligence as an Organizational Approach for Service Designers.”</a> <em>Touchpoint: The Journal of Service Design</em>, Volume 12, Issue 2, 2021.</p><p>Törmänen, Juha, Raimo P. Hämäläinen, and Esa Saarinen. <a href="https://onlinelibrary.wiley.com/doi/full/10.1002/hrdq.21455">“On the Systems Intelligence of a Learning Organization: Introducing a New Measure.”</a> <em>Human Resource Development Quarterly</em>, 2022.</p><p><em>(This article draws on publicly available sources and industry research. No proprietary or confidential information has been used. AI tools were used selectively to support research, editing, and visual development; the arguments and perspectives expressed are my own.)</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5128d573bc2c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The System is the Design.]]></title>
            <link>https://medium.com/@GregLakloufi/the-system-is-the-design-f17ee8bebd41?source=rss-2fa4496fe40b------2</link>
            <guid isPermaLink="false">https://medium.com/p/f17ee8bebd41</guid>
            <category><![CDATA[strategic-design]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[design-systems]]></category>
            <category><![CDATA[systems-thinking]]></category>
            <category><![CDATA[aerospace]]></category>
            <dc:creator><![CDATA[Greg Lakloufi]]></dc:creator>
            <pubDate>Wed, 26 Aug 2026 22:46:35 GMT</pubDate>
            <atom:updated>2026-09-14T17:47:08.601Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LZiSUegZ0HvVEKsSOzRauQ.png" /></figure><h4>What aerospace can teach us about the next evolution of strategic design.</h4><p>For decades, design has steadily expanded its field of view.</p><p>We began by designing objects. Then interfaces. Then experiences. Then services. Eventually, we began applying design to strategies, organizations, and transformations.</p><p>Each expansion happened for essentially the same reason: We discovered that the thing we were designing could no longer be understood independently from the world around it.</p><p>A product existed within an experience. An experience existed within a service. A service existed within an organization. And an organization existed within an increasingly complex network of technologies, markets, policies, people, infrastructure, incentives, culture, and behavior.</p><p>Our methods expanded because our understanding of the problem expanded.</p><p>But somewhere along the way, the complexity surrounding organizations began expanding faster than our methods did.</p><p>Today, we improve customer journeys. We redesign employee experiences. We modernize technology. We introduce AI. We restructure operating models. We optimize supply chains. We transform contact centers.</p><p>Each initiative may be entirely rational on its own.</p><p>And yet organizations frequently continue producing many of the same problems. Perhaps the problem is not that we need to design each of these things better. Perhaps we are designing at the wrong level.</p><p>The next evolution of design may not be another specialization of experience.</p><p>It may be learning how to see, and intelligently intervene in, <strong>the system producing the experience in the first place.</strong></p><p>I have increasingly come to think of that capability as <strong>Systems Intelligence</strong>.</p><p>For decades, designers have expanded the boundaries of what we consider design. The object became the interaction. The interaction became the experience. The experience became the service. The service became the organization.</p><p>Perhaps the next expansion requires a more fundamental realization:</p><p><strong>The system IS the design.</strong></p><p>And somewhat unexpectedly, one of the places I have begun looking for clues about what that means is aerospace.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-E5jiO2Y-O7fMCM8zC4A3w.png" /></figure><h3>We Are Getting Better at Solving the Wrong Problem</h3><p>Consider a large modern enterprise.</p><p>A customer encounters a problem and calls a contact center.</p><p>From the organization’s perspective, this becomes a contact-center problem. Call volume is too high. Handle times are too long. Agents need better tools. Self-service isn’t working. AI might automate some interactions.</p><p>So the organization launches a contact-center transformation.</p><p>But why did the customer call in the first place?</p><p>Perhaps the website communicated something incorrectly. Perhaps inventory data was inaccurate. Perhaps marketing created an expectation the operation couldn’t fulfill. Perhaps an employee didn’t have access to the right information. Perhaps a policy made perfect sense to Finance but created confusion for customers. Perhaps the technology worked exactly as designed but connected poorly with another system. Perhaps several independently reasonable decisions interacted to produce an unreasonable outcome.</p><p>The contact center didn’t create the problem. <br><strong>It received the consequence of the system.</strong></p><blockquote><strong>“Optimizing the contact center without understanding the system producing the calls may simply make the organization more efficient at absorbing failure.”</strong></blockquote><p>The same pattern appears everywhere.</p><p>Hospitals optimize clinical workflows while patients struggle across the larger care ecosystem. Banks redesign digital experiences while policies and legacy systems continue creating friction. Manufacturers automate processes without reconsidering the operating environment surrounding them. Organizations deploy AI into workflows that were poorly designed before AI arrived.</p><p>We keep encountering systemic problems and responding with localized solutions.</p><p>And then we are surprised when the problem moves somewhere else.</p><h3>Local Optimization Can Create Systemic Dysfunction</h3><p>This is one of the central lessons of complex systems.</p><p>Improving the performance of individual components does not necessarily improve the performance of the whole.</p><p>Sometimes it does the opposite.</p><p>A business unit optimizes its economics and transfers cost downstream. A digital team reduces friction for customers and unintentionally increases operational burden for employees. Automation removes human effort from one process while creating exceptions requiring greater human intervention somewhere else. Procurement reduces vendor costs while increasing maintenance complexity. Marketing increases demand while operations struggle to fulfill it.</p><p>Each function can demonstrate success against its own metrics.</p><p><strong>The system can still perform worse.</strong></p><p>This is what makes complexity fundamentally different from complicatedness.</p><p>A complicated object may contain thousands of parts and still behave predictably.</p><p>A complex system contains interactions capable of producing outcomes that cannot be understood simply by examining the parts independently.</p><p>This is precisely the territory Olivier de Weck, Daniel Roos, and Christopher Magee explore in <em>Engineering Systems</em>. In describing the premise behind the book, de Weck defines an engineering system as one containing both technical and social complexity:<em>“…both of these complexities are intertwined in a way that makes them inseparable.”</em></p><p><strong>Inseparable.</strong></p><p>That word matters.</p><p>The technical system and the human system are not two problems waiting to be independently solved and integrated later.</p><p><strong>Their interaction is the problem.</strong></p><p>That could describe a transportation network. It could describe an aerospace program. It could also describe a hospital, bank, retailer, government agency, or multinational enterprise. The boundary between technical systems and organizational systems is becoming increasingly difficult to draw.</p><p>And that should matter enormously to designers.</p><h3>Aerospace Learned This Lesson a Long Time Ago</h3><p>An aircraft is an extraordinary object.</p><p>But it is also not really an object at all.</p><p>It is the visible expression of an enormous collection of interacting decisions and constraints.</p><p>Aerodynamics affects structure. Structure affects weight. Weight affects propulsion. Propulsion affects range. Range affects fuel. Fuel affects weight again. Materials influence manufacturing. Manufacturing affects cost and maintainability. Human physiology affects cockpit design. Mission requirements influence nearly everything.</p><p>Change one variable and consequences propagate through the system.</p><p>That is one reason aerospace design provides such a useful lens through which to examine design and complexity.</p><p>Anthony Springer’s <em>Aerospace Design: Aircraft, Spacecraft, and the Art of Modern Flight</em> examines aerospace not simply as engineering, but through the relationships between form, function, experimentation, technological possibility, and performance.</p><p>In his contribution to the book, aerospace historian Michael H. Gorn makes an important observation: <em>“Flight research has offered moments of realization found in no other method.”</em></p><p>The shape of an aircraft is not simply the expression of a designer’s intent. It is an <strong>emergent answer to a system of constraints</strong>.</p><p>But aerospace learned something else as well: even sophisticated models cannot reveal everything about how a complex system will behave.</p><p>At some point, the system has to encounter reality:<br>&gt; <strong>Design. Test. Observe. Learn. Adapt.</strong></p><p>Some knowledge only emerges through that loop.</p><p>And the aircraft itself is only one component of something much larger: pilots, maintainers, bases, communications, logistics, training, weather, regulation, command structures, data, supply chains, intelligence, software, and increasingly autonomous technologies.</p><p>NASA’s definition of a system reflects this broader view. A system includes not merely hardware and software, but equipment, facilities, personnel, processes, and procedures. Much of its value exists not simply in those elements individually, but in the relationships between them.</p><p>That changes the unit of design.</p><p>The objective is no longer to make every component individually excellent.</p><p>The objective is to create a <strong>coherent whole capable of accomplishing a mission</strong>.</p><p>Aerospace did not have the luxury of pretending these things could be separated. A spacecraft cannot be 99 percent integrated. A flight-control system cannot ignore human factors because another team owns the cockpit. Propulsion cannot optimize itself independently of weight. Mission planners cannot assume logistics will somehow resolve themselves downstream.</p><p>A component can perform perfectly and still contribute to mission failure.<br><strong>The system has to work as a system.</strong></p><p>Organizations, however, still frequently structure transformation as though technology, people, operations, customers, strategy, and policy were separable domains.</p><p>We assign each to a function. We give each function metrics. We create workstreams. We appoint owners. Then we attempt to integrate everything later.</p><p>Aerospace suggests another possibility:</p><p><strong>&gt; Integration is not the final phase of design. Integration is the design problem.</strong></p><p>And increasingly, aerospace is taking that logic even further.<br>The aircraft itself is no longer the ultimate unit of capability.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*C6TLv9L1CZC2ulUzI3sAhA.png" /></figure><h3>The Aircraft Is Becoming a Node</h3><p>Perhaps the clearest evidence of this shift is the changing way modern combat aircraft themselves are understood.</p><p>For much of aviation history, the aircraft was naturally treated as the primary unit of capability:</p><ul><li>How fast could it fly?</li><li>How high could it climb?</li><li>What could it carry?</li><li>How maneuverable was it?</li><li>Could it outperform an opposing aircraft?</li></ul><p>Those characteristics still matter.</p><p>But they increasingly describe only part of what a modern aircraft actually does.</p><p>As early as the development of the F-22, U.S. Air Force leaders were arguing that traditional labels such as “fighter” were becoming inadequate. The aircraft could penetrate contested airspace, collect information, fuse sensor data, contribute to surveillance and reconnaissance, and share awareness across a larger force.</p><p>It was not simply an aircraft carrying sensors.<br>It was becoming <strong>a sensor that happened to fly</strong>.</p><p>The F-35 takes that idea considerably further.</p><p>Its value is not limited to what the aircraft itself can see, engage, or destroy. Its sensors collect enormous amounts of information. Onboard systems fuse that information into an integrated picture. The aircraft can then contribute what it knows to other aircraft, ground forces, ships, command systems, and other elements of a joint force.</p><p>Lockheed Martin describes the F-35 explicitly as <strong>“the most advanced node”</strong> within a network-centric architecture connecting air, land, sea, space, and cyber operations.</p><p>That language represents a profound change in the idea of the aircraft itself.</p><p>The airplane is no longer only the product.<br>It is a <strong>node within a system of systems</strong>.</p><p>Its capability emerges partly from what it can do independently, but increasingly from what it can sense, contribute, receive, and enable across the larger network.</p><p>One platform can detect something another cannot see. Another system can provide information the aircraft could not have generated itself. A sensor can cue a shooter. A platform designed for one role can create advantage somewhere entirely different in the system.</p><p>The value moves from the object toward the <strong>relationships among objects</strong>.</p><p>And the next generation of airpower is making that relationship even more explicit.</p><p>The Air Force’s Collaborative Combat Aircraft (CCA) program introduces autonomous, uncrewed aircraft designed from the outset to operate alongside crewed platforms as part of a human-machine team. These aircraft are not conceived primarily as standalone replacements for traditional fighters. Their purpose is inherently relational: extending the reach, survivability, sensing, and effectiveness of the larger force.</p><p>The Air Force itself describes CCA as a critical component of a larger <strong>“integrated system-of-systems”—</strong> a network capable of distributing sensing, striking, shielding, and decision advantage across multiple platforms.</p><p>This represents another important shift.</p><p>The F-35 demonstrates how an aircraft can become an intelligent node within a network. CCA demonstrates what becomes possible when we begin designing the nodes specifically FOR the network.</p><p>Capability is no longer concentrated entirely in one exquisite platform. It can increasingly be distributed across crewed and uncrewed systems, each contributing different capabilities to a larger mission.</p><p>The unit of design moves again: <strong>From platform &gt; to node &gt; to network &gt; to system of systems.</strong></p><p>And buried inside that evolution is an important lesson for design.<br>Traditional design tends to ask:<br><strong>• How do we make the node better?<br></strong>Systemic design asks a fundamentally different question:<br><strong>• What combination of nodes makes the system better?</strong></p><p>That shift sounds subtle. <strong>It isn’t.</strong></p><p>It moves the object of design from the individual component to the relationships, interactions, and collective capabilities that emerge between components.</p><p>The question is no longer simply:<br><strong>• How capable is this aircraft?<br></strong>It becomes:<br><strong>• How does this aircraft make the entire system more capable?</strong></p><p>And that may be one of the most consequential lessons aerospace offers organizational design. Because modern organizations are beginning to face the same transition.</p><p>A contact center is not simply a contact center. A store is not simply a store. An AI agent is not simply an AI agent. A digital channel is not simply a digital channel.</p><p>Each is a node that senses, receives information, makes decisions, creates consequences, and influences the performance of the larger system.</p><p>Perhaps the strategic question should therefore change from:<br><strong>• How do we optimize this thing?<br></strong>to:<br><strong>• What role should this node play in making the entire system more intelligent, adaptive, and effective?</strong></p><p>That is a very different design problem.</p><h3>Human-Centered Design Was Necessary. It Is No Longer Sufficient.</h3><p>This is not an argument against human-centered design.</p><p>Quite the opposite.</p><p>Human-centered design represented an essential correction to a world that too often designed technology, products, and processes around organizational requirements rather than human needs.</p><p>It taught organizations to observe. To listen. To empathize. To understand context. To design with people rather than merely for them.</p><p>Those capabilities remain fundamental.</p><p>But the complexity surrounding those humans has changed dramatically.</p><p>The person at the center of an experience now exists inside interconnected technological, organizational, economic, regulatory, and increasingly algorithmic systems.</p><p>Understanding that person’s needs does not necessarily tell us how the system producing their experience needs to change.</p><p>And optimizing the experience of one participant may create unintended consequences for another.</p><p>Customers want immediate fulfillment. Employees need manageable workloads. Businesses need sustainable economics. Algorithms optimize against defined objectives. Regulators impose constraints. Technology creates possibilities and dependencies. Supply chains introduce uncertainty. Leaders operate within organizational and financial realities.</p><p>None of these perspectives is inherently wrong.<br>The challenge is understanding <strong>what happens between them</strong>.</p><p>Human-centered design therefore does not need to disappear.<br>Its aperture needs to widen.</p><p><strong>The human remains essential. <br>But the system becomes the unit of inquiry.</strong></p><h3>Design Has Always Widened Its Aperture</h3><p>There is a useful way to understand how we arrived here.</p><ul><li><strong>Industrial Design</strong> helped us shape objects.</li><li><strong>Interaction Design</strong> helped us shape relationships between humans and technology.</li><li><strong>Experience Design</strong> expanded our view across interactions.</li><li><strong>Service Design</strong> connected experiences across channels, actors, processes, and time.</li><li><strong>Strategic Design</strong> moved further upstream, applying design to business decisions, organizations, and transformation.</li></ul><p>Each evolution expanded what designers believed they were responsible for understanding.</p><p>Systems Intelligence may represent another widening.<br>Not simply:<br>• What experience should we create?<br>Or:<br>• What service should we redesign?<br>But:<br><strong>• What system is producing the outcome we are observing?</strong></p><p>Who participates in it? What relationships connect them? Where does information move, or fail to move? Where are decisions made? What incentives shape behavior? What assumptions have become invisible? What constraints are fixed, and which merely appear to be? Where does the system adapt? Where does it resist? What feedback loops reinforce its current behavior?</p><p>And critically:<br><strong>• If we intervene here, what happens somewhere else?</strong></p><p>These are design questions.<br>But they are also systems questions.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_T2O4mLBm6qOEb4TKCC-mg.png" /></figure><h3>Skunk Works Discovered That the Organization Is Part of the Technology</h3><p>Aerospace offers another lesson about systems, but this one isn’t found in the aircraft. It’s found in the organization that made them possible.</p><p>Lockheed’s legendary Skunk Works is remembered for extraordinary machines such as the U-2 and SR-71, and for the lineage that later produced stealth aircraft including the F-117, F-22, and F-35.</p><p>At the center of its creation was Clarence “Kelly” Johnson, one of the twentieth century’s most influential aircraft designers, who founded Skunk Works and shepherded some forty military and civilian aircraft into production. Johnson understood that extraordinary aircraft required more than extraordinary engineering.</p><p>They required a different kind of organization.</p><p>He didn’t simply assemble brilliant engineers and ask them to innovate harder.</p><p><strong>He changed the conditions surrounding them.</strong></p><p>Small teams. Significant authority. Direct communication. Reduced bureaucracy. Close relationships between engineers and fabricators. Rapid decisions. Individual accountability. Experimentation. And an almost obsessive focus on the mission.</p><p>Johnson recognized that the organizational environment surrounding engineers would directly influence what those engineers were capable of producing.</p><p>Josh Dean’s <em>The Impossible Factory</em> reveals another important dimension of Johnson’s thinking.</p><p>At the time, aircraft designers could develop a prototype and effectively hand it downstream for someone else to test. Johnson rejected that separation between the people designing the aircraft and the people discovering how it actually behaved.</p><p>As Dean writes: <em>“It was the designer’s responsibility to prove a plane did what it was supposed to do.”</em></p><p>Engineers and test pilots, Johnson insisted, <strong>“must live together.”</strong></p><p>That principle is deceptively profound. Johnson wasn’t merely improving collaboration. <strong>He was protecting the feedback loop between design and consequence. </strong>The people designing the system remained connected to what happened when their decisions encountered reality.</p><p>And this wasn’t merely an informal Skunk Works habit. Johnson eventually encoded the principle into his operating rules: <em>“The contractor must be delegated the authority to test his final product in flight.”</em></p><p>But the warning that followed may be even more consequential: <em>“If he doesn’t, he rapidly loses his competency to design other vehicles.”</em></p><p>That is a remarkably modern observation.</p><p>Johnson understood that separating designers from the consequences of their decisions didn’t merely weaken the product.</p><p><strong>Eventually, it weakened the designers.</strong></p><p>Separate people from the behavior of the system long enough, and they begin to lose their understanding of the system itself.</p><p>The breakthrough was not merely technological.<br><strong>The system producing the technology had been redesigned.</strong></p><p>Johnson repeatedly examined where traditional engineering hierarchies created delay: slow communication, unclear responsibility, too many decision-makers, too many organizational layers.</p><p><strong>Skunk Works was not only an engineering achievement. It was an institutional design achievement.</strong></p><p>Authority, proximity, trust, feedback, and mission were deliberately configured so the organization could learn at the speed of the problem. The aircraft were extraordinary. But the operating system that produced them may have been the more consequential innovation.</p><p>The lesson is remarkably contemporary.</p><p>Organizations routinely say they want innovation while preserving the structures that prevent innovation from occurring.</p><p>They create innovation labs inside governance systems designed primarily to minimize risk. They ask employees to experiment while rewarding predictability. They introduce agile teams while retaining slow decision structures. They deploy AI while preserving workflows created for a pre-AI world. Then they wonder why transformation stalls.</p><p>The problem may not be the people.<br><strong>The organization may be behaving exactly as its system encourages it to behave.</strong></p><h3>Systems Produce Behavior</h3><p>This may be one of the most important ideas for leaders — and designers — to internalize.</p><p>When the same problem repeatedly appears across different people, teams, technologies, or initiatives, we should question whether the problem actually resides in any of them.</p><p>Repeated outcomes are often evidence of underlying structures: Incentives. Information flows. Governance. Dependencies. Decision rights. Policies. Feedback loops. Cultural norms. Technology architecture. Organizational boundaries…</p><p>These structures quietly shape behavior.</p><p>Yet many are invisible in traditional organizational charts, process maps, and customer journeys.</p><p>So organizations repeatedly treat their manifestations as isolated events: A customer complaint becomes a CX problem. Employee frustration becomes an EX problem. A broken workflow becomes a technology problem. Slow execution becomes an operating-model problem. Low adoption becomes a change-management problem.</p><p>Each problem is handed to a different team.<br>Each team does good work.</p><p>And nobody owns the relationships between them.</p><p><strong>Increasingly, the most consequential problems may live in the spaces between organizational domains.</strong></p><p>This tension becomes particularly visible in defense, where individual organizations can successfully accomplish their assigned objectives while the larger system still struggles to create sustained advantage.</p><p>As Colonel Jason Trew, PhD (USAF Ret.), former Commandant and Dean of the <a href="https://www.airuniversity.af.edu/SAASS/">School of Advanced Air &amp; Space Studies</a>, puts it:</p><blockquote>“The military is highly attuned to winning — and when you look at the small, bounded problems individual offices focus on, it can genuinely look like victory. But strategy is singular; that is, it is about advantage across the whole system over time. That requires systemic design, not just systematic modifications.”</blockquote><p>His distinction between <strong>systematic modification </strong>and<strong> systemic design</strong> is an important one.</p><p>The former improves parts of the existing system.<br>The latter asks whether those parts, their relationships, and perhaps even the boundaries of the system itself need to change.</p><p>And that points toward something design needs to become much better at seeing.</p><p>Not simply the nodes.<br><strong>The relationships between them.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4Kqx8JFfGlI1cEQxriV7Hw.png" /></figure><h3>AI Will Expose How Fragmented Our Systems Really Are</h3><p>And now this problem is becoming urgent.</p><p>For years, organizations could tolerate enormous fragmentation because humans served as connective tissue: Employees reconciled conflicting information. Managers interpreted ambiguous policies. Customer-service agents worked around broken systems. People copied information between applications. Experienced employees knew whom to call when the official process failed. Teams invented unofficial processes to compensate for official ones. <strong>Human beings quietly held fragmented organizations together.</strong></p><p>AI changes that equation.</p><p>As intelligent agents begin participating in decisions, workflows, customer interactions, knowledge management, and operations, organizations will increasingly automate not simply tasks, but <strong>relationships between tasks</strong>.</p><p>That creates extraordinary opportunity.<br>It also introduces a dangerous possibility.</p><p>Poorly understood systems can begin operating faster.</p><p>AI inserted into a dysfunctional system does not automatically create an intelligent organization.</p><p>It may simply create a <strong>faster dysfunctional system</strong>.</p><p>Before asking what AI can automate, organizations may therefore need to ask a more fundamental question:</p><ul><li><strong>What system are we automating?</strong></li></ul><p>And perhaps an even more important one:</p><ul><li><strong>What system should exist instead?</strong></li></ul><h3>Systems Intelligence meets Strategic Design</h3><p>None of this suggests that we have discovered systems for the first time.</p><p>Far from it.</p><ul><li><strong>Systems thinking</strong> provides a rich body of knowledge for understanding relationships, feedback, emergence, and complexity.</li><li><strong>Systems engineering</strong> provides rigorous methods for designing and integrating complex technical and sociotechnical systems.</li><li><strong>Design</strong> provides powerful methods for understanding human needs, reframing problems, imagining alternatives, experimenting, and making possibilities tangible.</li><li><strong>Strategy</strong> helps organizations make choices under uncertainty.</li><li><strong>Organizational design </strong>shapes the structures through which work happens.</li><li><strong>Futures thinking</strong> helps us consider multiple possible environments rather than assuming continuity.</li></ul><p>None of these disciplines needs replacing.</p><p><strong>The opportunity lies in connecting them.</strong></p><p>Systems Intelligence is not a new term. Raimo Hämäläinen and Esa Saarinen introduced the concept more than two decades ago to describe intelligent human behavior within complex systems. Their work explores how people perceive their participation in a larger whole, recognize feedback, and act productively even when the system cannot be fully known or controlled.</p><p>That human-centered foundation matters. The aerospace challenge extends it into a larger design question: how do we intentionally shape the structures, technologies, relationships, governance, and learning mechanisms through which an institution can behave intelligently as a whole?</p><p>This is where Systems Intelligence meets Strategic Design.</p><p><strong>Strategic Design &amp; Systems Intelligence, or SDSI, is the applied practice of perceiving, designing, and deliberately intervening in complex human, organizational, and technological systems so that institutions can sense, interpret, decide, intervene, and learn as coherent adaptive wholes.</strong></p><p>It requires seeing beyond individual experiences toward the conditions producing them. Beyond individual problems toward recurring patterns. Beyond organizational boundaries toward relationships. Beyond static maps toward dynamic behavior. Beyond solutions toward consequences. Beyond optimization toward adaptation. Beyond prediction toward preparedness. And perhaps most importantly:</p><p><strong>Beyond designing individual things toward designing the conditions through which better outcomes can emerge.</strong></p><p>SDSI is not systems thinking with a new label.</p><p>Systems thinking helps us understand complexity. Systems Intelligence helps people perceive and act intelligently within it. Strategic Design helps us deliberately reshape the conditions producing that complexity.</p><p>Together, they create an applied practice for designing more intelligent and adaptive institutions.</p><p>That means combining analytical understanding with human observation, strategic judgment, systems modeling, visualization, experimentation, prototyping, and continuous learning.</p><p>It also requires accepting something designers and business leaders are not always comfortable accepting: Complex systems cannot always be controlled.</p><p>Sometimes they must be probed. Observed. Influenced. Adapted. And learned from. SDSI cannot be practiced at a distance from the system.</p><p>The objective is not to eliminate complexity.<br><strong>It is to make complexity intelligible enough to act upon.</strong></p><h3>The Designer’s Aperture Must Widen Again</h3><p>Design has always evolved by expanding what designers believe they are responsible for.</p><p>The chair became the room. The interface became the journey. The journey became the service. The service became the organization. Now the organization is becoming part of something larger still.</p><p>Markets. Technologies. AI. Infrastructure. Policy. Culture. Geopolitics. Supply chains. Human behavior. Climate. Security...</p><p>These forces increasingly collide inside every meaningful transformation.</p><p>We cannot predict all of their interactions. We cannot control all of their outcomes. And perhaps trying to do so is itself part of the problem.</p><p>But we can become considerably better at <strong>seeing the systems we are operating within.</strong></p><p>We can map relationships instead of merely documenting components. We can identify feedback instead of merely measuring performance. We can expose assumptions instead of allowing them to become constraints. We can experiment with interventions rather than betting everything on predetermined solutions. We can examine second-order consequences before they become tomorrow’s transformation program. We can design organizations capable of learning and adapting as conditions change.</p><p>And perhaps this is the most important lesson aerospace offers us.</p><p>The extraordinary machines we admire are not extraordinary because someone perfectly designed every individual component.</p><p>They work because thousands of components, people, decisions, constraints, and relationships were brought together into a coherent system capable of accomplishing a mission.</p><p>Increasingly, organizations face the same challenge.</p><p>The future of design may therefore depend less on our ability to design better things.</p><p>It may depend on our ability to understand <strong>what those things belong to</strong>.</p><p>Because the next great design problem isn’t the experience.<br><strong>It is the system producing it.</strong></p><p>And if that is true, our responsibility as designers and strategists must expand once again.</p><p>We cannot simply design better outcomes.<br><strong>We must become capable of designing the systems that produce them.</strong></p><p><em>This article examines the system from the perspective of capability. In</em> <a href="https://medium.com/@GregLakloufi/artificial-intelligence-needs-systems-intelligence-5128d573bc2c">Artificial Intelligence Needs Systems Intelligence</a>, <em>I apply the same systems-of-systems logic to AI governance and introduce the SDSI operating model for institutions capable of sensing, interpreting, deciding, intervening, and learning as adaptive wholes.</em></p><p>References:<br>Hämäläinen, Raimo P., and Esa Saarinen, editors. <em>Systems Intelligence: </em><a href="https://research.aalto.fi/en/publications/systems-intelligence-discovering-a-hidden-competence-in-human-act/"><em>Discovering a Hidden Competence in Human Action and Organizational Life.</em></a> Helsinki University of Technology, 2004.</p><p>Harviainen, J. Tuomas, Raimo P. Hämäläinen, and Esa Saarinen. <a href="https://research.aalto.fi/en/publications/system-intelligence-as-an-organizational-approach-for-service-des/">“Systems Intelligence as an Organizational Approach for Service Designers.”</a> <em>Touchpoint: The Journal of Service Design</em>, Volume 12, Issue 2, 2021.</p><p><em>(This article draws on publicly available sources and industry research. No proprietary or confidential information has been used. AI tools were used selectively to support research, editing, and visual development; the arguments and perspectives expressed are my own.)</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f17ee8bebd41" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Age of Systems: Why Products and Services Are No Longer Enough]]></title>
            <link>https://medium.com/@GregLakloufi/the-age-of-systems-why-products-and-services-are-no-longer-enough-ec77e50d947f?source=rss-2fa4496fe40b------2</link>
            <guid isPermaLink="false">https://medium.com/p/ec77e50d947f</guid>
            <category><![CDATA[systems-thinking]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[strategic-design]]></category>
            <category><![CDATA[business-transformation]]></category>
            <category><![CDATA[innovation]]></category>
            <dc:creator><![CDATA[Greg Lakloufi]]></dc:creator>
            <pubDate>Wed, 10 Jun 2026 15:44:10 GMT</pubDate>
            <atom:updated>2026-09-16T19:28:24.544Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wKkthsPlhSBHLcpz6ojZeQ.jpeg" /></figure><h4>Products can be copied. Services can be replicated. Systems determine what endures.</h4><p>There is a strange contradiction taking shape inside modern business.</p><p>Organizations have never been better equipped to improve things. They have more data, more technology, more specialized talent, more design methods, more agile rituals, more customer research, more innovation programs, and more transformation budgets than at any point in history. And now with AI and seemingly unlimited data, they can test ideas faster, launch products faster, personalize experiences faster, and measure behavior with astonishing precision.</p><p>And yet many organizations feel more fragile than ever.</p><p>This fragility does not always appear as dramatic failure. More often, it appears as disappointment. But why?<br>A product launches successfully but never creates the expected business impact. A service is redesigned thoughtfully but remains difficult to deliver consistently. A digital transformation meets its milestones but fails to change how the organization actually works. A new technology creates excitement in pilots but struggles to scale beyond controlled environments.</p><p>At first, these seem like familiar execution problems. The roadmap was too ambitious. The operating model was unclear. The governance was too slow. The adoption plan was weak. The technology stack was fragmented. Sometimes those explanations are true. But increasingly, they feel incomplete.</p><p>Because in many cases, the product did not fail.<br>The service did not fail.<br>The experience did not fail.</p><p>The system supporting them did.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*K4YIamDl8t_6G_ziOOm1fQ.jpeg" /></figure><h3>The Illusion of Progress</h3><p>Over the past few years, I have seen versions of this pattern surface across industries. In healthcare, organizations invest heavily in patient-centered experiences, virtual care models, and digital access points, only to discover that clinical workflows, reimbursement structures, staffing constraints, and legacy systems continue to shape the actual outcome far more than the interface ever could. In financial services, firms modernize digital onboarding and customer self-service, yet the experience remains constrained by fragmented ownership, regulatory interpretation, risk processes, and decades of operational complexity. In retail, brands launch new products with great promise, only to discover that supply chains, fulfillment networks, store operations, and data infrastructure determine whether the promise can be kept.</p><p>None of these organizations are failing because they lack ambition. In many cases, the ambition is clear. The talent is strong. The strategy is reasonable. The design work is thoughtful. The technology is capable.</p><p>The issue is that the work is often aimed at the most visible part of the problem rather than the system producing it.</p><p>This is one of the reasons transformation can feel so frustrating. Organizations are able to improve components while leaving the deeper dynamics untouched. They can make a journey more elegant while the operating model remains fragmented. They can make a product more intelligent while the governance surrounding it remains slow. They can make a service more human while the incentives behind it continue to reward the wrong behavior.</p><p>The improvement is real.<br>It is just not enough.</p><p>Increasingly, organizations are discovering that improving a component without understanding the incentives surrounding it can produce unexpected outcomes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*mJKJRwzI5noJffnr5IaLVw.jpeg" /></figure><p>Recently, <a href="https://www.businessinsider.com/amazon-ai-leaderboard-tokenmaxxing-2026-5">Amazon</a> encountered a revealing challenge as employees experimented with an informal, employee-created leaderboard called KiroRank, which ranked participants according to the volume of tokens consumed through AI tools.</p><p>The tracker was not an official performance-management system. But once token consumption became visible as a measure of status, some employees reportedly began using AI unnecessarily to improve their position, a behavior described as “tokenmaxxing.” Amazon subsequently discontinued the leaderboard.</p><p>What appeared to measure AI adoption actually measured AI activity, and once that activity became a status signal, behavior adapted to the metric. What appeared to be an AI adoption problem was actually an incentives problem.</p><p>The distinction matters. The organization did not intentionally design people to waste tokens. But the metric produced the behavior it rewarded.</p><p>What appeared to be an AI adoption problem was actually an incentives problem.</p><p>That is the uncomfortable truth many organizations are now confronting. The tools that helped them succeed in one era are not always sufficient for the realities of the next.</p><h3>The World We Designed For No Longer Exists</h3><p>For much of the twentieth century, the product was the dominant unit of value. Organizations that could design, manufacture, distribute, and scale better products created enormous advantage. The logic was clear because the world rewarded it. Better products meant better performance, greater efficiency, stronger differentiation, and larger markets.</p><p>Then, as products became easier to copy and markets became more saturated, value began shifting toward services and experiences. The question was no longer only what a company made, but how people experienced what it made. A hotel room became hospitality. A bank account became financial confidence. A phone became a gateway into a digital life. The rise of User Experience, Customer Experience, and Service Design reflected this important evolution. These disciplines helped organizations see people less as transactions and more as participants in journeys, relationships, and ecosystems.</p><p>Service Design, in particular, gave organizations a more expansive way to think about value. It helped reveal the connections between touchpoints, channels, frontstage interactions, backstage operations, and human outcomes. It challenged the narrowness of product-centric thinking and brought a much-needed level of empathy and orchestration into business.</p><p>But every discipline is shaped by the problems of its time. Product strategy emerged when the central challenge was making and scaling things. Service strategy emerged when the central challenge was orchestrating experiences across channels and moments. Both remain valuable. Neither disappears. But the world around them has changed, yet again.</p><p>The defining challenge of the 21st century is not simply making better products or delivering better services. It is building systems capable of adapting inside complexity.</p><p>This is a different kind of challenge.</p><p>It is less linear. Less contained. Less forgiving. It does not stay neatly inside a product team, a design team, a customer journey, a technology platform, or an operating model. It moves across them.</p><p>A supply chain disruption becomes a customer experience problem. A regulatory shift becomes a product strategy problem. A climate event becomes a logistics problem. An AI capability becomes an organizational design problem. A geopolitical conflict becomes a pricing, sourcing, staffing, and trust problem all at once.</p><p>The old boundaries no longer hold.</p><p>And when the boundaries no longer hold, the way we design has to change.</p><h3>The System as the New Competitive Advantage</h3><p>For much of the last century, competitive advantage was relatively easy to identify. Organizations that could design, manufacture, and distribute better products tended to win. As products became easier to replicate, advantage increasingly shifted toward services and experiences. Organizations that could create more meaningful relationships with customers often outperformed those that focused solely on what they produced.</p><p>For a time, this logic held.</p><p>But as complexity increased, something else began to happen.</p><p>Somewhere along the way, the source of competitive advantage quietly moved beneath the surface. The organizations creating the most enduring value were no longer winning because of a single product or even a superior service. They were winning because of their ability to connect products, services, people, technology, operations, intelligence, and decision-making into coherent systems capable of learning and adapting over time.</p><p>This shift helps explain why so many transformation efforts feel simultaneously successful and disappointing. Organizations continue improving the visible parts of the business because those are the parts they know how to manage. They redesign experiences, modernize platforms, launch products, and deploy new technologies. Yet many of the challenges they face originate not within those components, but within the relationships between them.</p><p>The work creates movement.<br>It simply doesn’t always create transformation.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tscgQb5vFIhDMSIdg6pAQg.jpeg" /></figure><h3>When Complexity Reveals the Cracks in the System</h3><p>The pandemic made this visible in a way few business events ever have. At first, many organizations tried to understand the disruption through the lens closest to them. Some saw a supply chain crisis. Others saw a workforce crisis. Others saw a customer behavior crisis. Others saw a technology acceleration crisis.</p><p>They were all right.</p><p>And that was precisely the point.</p><p>The disruption was not one thing. It was a system under stress. A health crisis became an economic crisis. An economic crisis became a labor crisis. A labor crisis became a supply chain crisis. A supply chain crisis became an inflation crisis. Every part influenced the next, and every response created new consequences somewhere else.</p><p>This is how systems behave. They do not respect the categories organizations use to manage them. They do not care where one department ends and another begins. They do not move according to quarterly planning cycles. They react, adapt, amplify, and sometimes break in ways that are difficult to predict from any single vantage point.</p><p>The same dynamic is now unfolding with perhaps the most consequential technology shift of our time: artificial intelligence.</p><p>Many organizations are approaching AI as a technology initiative, which is understandable. They are evaluating tools, comparing models, launching pilots, and building governance frameworks. All of that is necessary.</p><p>But AI is not only a toolset. It is becoming a new layer of organizational intelligence. It changes how knowledge is created, how decisions are made, how work is distributed, how expertise is accessed, how services are delivered, and how customers interact with organizations. It challenges assumptions about talent, process, governance, accountability, and value creation.</p><p>In other words, AI does not simply enter the system.<br>It changes the behavior of the system.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ySEHoh0MgM3MlJ8IQ_zxyw.jpeg" /></figure><p>Starbucks recently provided a useful example of how difficult this can be in practice. The company deployed an AI-enabled inventory-counting system across more than 11,000 North American stores to improve visibility, reduce shortages, and make inventory work more efficient.</p><p>In practice, the system reportedly miscounted and mislabeled products, sometimes confusing similar milk types or failing to recognize items altogether. Employees had to correct the resulting errors, introducing new work into the process the technology was intended to simplify. Starbucks discontinued the system approximately nine months after deployment.</p><p>The technology did not operate independently of the physical environment, product variations, store organization, and people responsible for correcting its output.</p><p>What appeared to be an inventory-counting problem revealed itself as a systems problem.</p><p>The same dynamic has surfaced in entirely different industries.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-Y_aubsplb7l4bBsd7Ro8Q.jpeg" /></figure><p>Zillow’s now-famous <a href="https://www.nytimes.com/2021/11/02/business/zillow-houses-flip.html">Zillow Offers</a> initiative was built on a compelling premise. Armed with vast amounts of market data and increasingly sophisticated predictive models, the company believed it could accurately purchase, renovate, and resell homes at scale. On paper, the logic appeared sound. The models could estimate value. The technology appeared capable. The market opportunity was enormous.</p><p>What Zillow ultimately discovered was that housing is not simply a pricing problem.</p><p>It is a system.</p><p>Home values are influenced by labor availability, renovation capacity, supply chains, regional market dynamics, interest rates, consumer behavior, and countless other interconnected variables. Zillow could estimate the value of an individual property. What it could not reliably predict at scale was how price volatility, renovation capacity, labor, supply chains, holding periods, and regional market conditions would interact across thousands of simultaneous transactions.</p><p>The issue was not that Zillow lacked data.<br>Nor was it that the technology was unsophisticated.</p><p>The challenge was that the housing market behaved less like a pricing model and more like a living system.</p><p>What appeared to be a forecasting problem revealed itself as a systems problem.</p><p>And the pattern extends well beyond retail operations.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Gwgek31cvTLvg6DItrvYlg.jpeg" /></figure><p>A <a href="https://www.businessinsider.com/pizza-hut-ai-system-dragontail-lawsuit-franchisee-2026-5">Pizza Hut</a> franchisee has alleged a similar dynamic involving the company’s Dragontail delivery platform. In a lawsuit, Chaac Pizza Northeast claims the mandated system changed how third-party drivers responded to available delivery information, contributed to delayed pickups, and created cascading operational consequences across 111 restaurants. Pizza Hut has not conceded those claims, and the dispute remains an allegation rather than an established finding. If the allegations are substantiated, the case illustrates how optimizing one part of a delivery system can unintentionally disrupt others.</p><p>Again, the lesson was not simply that the technology failed.<br>The lesson was that systems often react in ways that individual components cannot predict.</p><p>These examples may appear unrelated. One involves incentives. Another inventory management. Another delivery logistics. Yet each reveals the same underlying truth. The challenge was not the technology alone. The challenge was understanding the system into which the technology was introduced.</p><p>Complexity did not create the problem.<br>It exposed it.</p><p>And once the cracks become visible, a deeper realization follows.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*B4RDMMWentZEoD7Y3mz2nA.jpeg" /></figure><h3>Systems Now Compete Against Systems</h3><p>This is why some of the most successful organizations of the modern era are so difficult to understand through traditional categories.</p><p>Apple is often described as a product company, and at one level that is true. The iPhone is one of the most successful products ever created. Its design, usability, and integration have shaped an entire generation of technology expectations.</p><p>But Apple’s advantage is not simply the iPhone.<br>If it were, competitors would have closed the gap long ago.</p><p>Hardware improves quickly. Features are copied. Interfaces converge. The visible product, however exceptional, is only part of the story.</p><p>What makes Apple resilient is the system around the product: hardware, software, services, payments, retail, developers, supply chains, subscriptions, identity, content, cloud infrastructure, and brand trust all reinforcing one another. In fact, in fiscal 2025, <strong>Apple’s services business generated approximately </strong><a href="https://investor.apple.com/"><strong>$109.2 billion</strong></a><strong> in revenue</strong>, exceeding the combined fiscal 2025 revenues of <a href="https://investors.nike.com/investors/news-events-and-reports/investor-news/investor-news-details/2025/NIKE-Inc--Reports-Fiscal-2025-Fourth-Quarter-and-Full-Year-Results/default.aspx">Nike ($46.3 billion)</a> and <a href="https://corporate.mcdonalds.com/content/dam/sites/corp/nfl/pdf/Investor%20Overview%20Deck_5.13.26.pdf">McDonald’s ($26.9 billion)</a><strong>. </strong>Hidden beneath the hardware is a vast ecosystem of subscriptions, payments, content, cloud services, and digital experiences that most customers rarely think about, yet collectively create one of the most powerful business engines in the world.</p><blockquote>The product captures the attention. The system captures the value. The customer experiences a device. The market competes with an ecosystem.</blockquote><p>Apple’s advantage does not reside in any one component. It emerges from the organization’s ability to coordinate the relationships among them.</p><p>This is what makes systems more defensible than products or services alone. A competitor can reproduce an individual feature. Reproducing the relationships, capabilities, operating structures, and accumulated learning surrounding it is far more difficult.</p><p>The same is true elsewhere.</p><p>Amazon is not simply a retailer. Its advantage lives in the relationship between marketplace, logistics, cloud infrastructure, data, fulfillment, payments, advertising, and increasingly AI.</p><p>Tesla is not simply an automotive company. Its position is shaped by software, charging infrastructure, data collection, manufacturing philosophy, energy systems, and brand mythology.</p><p>SpaceX is not simply building rockets. It redesigned the relationship between engineering, manufacturing, testing, failure, iteration, cost, and speed.</p><p>The visible artifact gets the attention.<br>The underlying system creates the advantage.</p><p>A system is not simply a larger collection of products and services. It is the network of relationships, dependencies, incentives, information flows, technologies, and feedback loops through which those elements produce value together.</p><p>This is the shift many organizations have not fully internalized. The primary unit of competition is changing. Products, services, and experiences remain important. But increasingly, they are expressions of larger systems rather than isolated sources of advantage.</p><p>A product can be copied. A service can be replicated. A feature can be matched. A journey can be redesigned.</p><p><strong>A resilient system is much harder to reproduce.</strong></p><p>That is why the next frontier of advantage will not come from asking only, “How do we make this product better?” or “How do we improve this service?”</p><p>Those questions still matter. But they are incomplete. The more important questions are becoming:</p><p>What system makes this product possible?<br>What conditions allow this service to perform?<br>What dependencies make this experience fragile?<br>What feedback loops shape behavior?<br>What incentives distort the outcome?<br>What forms of intelligence are missing?<br>Where does the system fail under pressure?</p><p>Those are not traditional product questions.<br>They are not traditional service questions.<br>They are systems questions.</p><p>And answering them requires a different way of seeing. Not a better view of the parts, but a deeper understanding of the whole.</p><p>The organizations that thrive in the decades ahead will still build great products and deliver exceptional services. The difference is that they will recognize both as outcomes of something deeper: the systems that enable them to learn, adapt, and create value continuously in an increasingly uncertain world. I explore the design implications of this transition more deeply in <a href="https://medium.com/@GregLakloufi/the-system-is-the-design-f17ee8bebd41"><em>The System Is the Design</em></a>, including what aerospace and modern systems-of-systems reveal about the expanding responsibility of design.</p><p>For many organizations, this realization challenges some of the very disciplines that helped them succeed over the past two decades. If systems are becoming the primary source of resilience and advantage, then the question is no longer how we design better experiences alone. It is how we design the systems that produce them.</p><h3>Designing Beyond the Journey</h3><p>This is where the conversation becomes uncomfortable, especially for those of us who have spent years advocating for Service Design as a more holistic way to create value.</p><p>Service Design is of course not obsolete. In many organizations, it is still underused, misunderstood, or reduced to journey maps and workshop artifacts. At its best, <strong>Service Design remains one of the most powerful disciplines for understanding how people, processes, technologies, and organizations interact to produce experiences.</strong></p><p>But the next era will ask more from it.</p><p>The risk is not that Service Design disappears. The risk is that it remains too attached to the problems it originally evolved to solve. Many service design methods were built to help organizations move beyond products and understand experiences across channels, moments, and touchpoints. That was essential work. But the challenges organizations now face increasingly extend beyond the orchestration of experiences.</p><p>They involve resilience, adaptability, governance, intelligence, ecosystem behavior, operational complexity, and the ability to sense and respond to change.</p><p>A beautifully designed service can still fail if the incentives behind it are misaligned. A seamless digital journey can still fail if the operating model cannot support it. A customer-centered vision can still fail if the organization is structured in ways that prevent collaboration. A brilliant AI-enabled experience can still create risk if governance, data quality, accountability, and human oversight are not designed as part of the system.</p><p>This does not diminish the value of Service Design. It expands its responsibility.</p><p>The discipline must move from designing services to designing the conditions in which services can survive, adapt, and create value. It must become less focused on the artifact of the journey and more focused on the behavior of the system. It must learn to operate not only at the level of experience, but at the level of organizational intelligence.</p><p>That is a meaningful evolution.<br>And it opens the door to something larger.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0_lx1BcTF07Id_ulvzCcqQ.jpeg" /></figure><h3>Where Systems Intelligence Meets Strategic Design</h3><p>Systems Intelligence did not begin with this article, nor did it begin with design.</p><p>Raimo Hämäläinen and Esa Saarinen <a href="https://research.aalto.fi/en/publications/systems-intelligence-discovering-a-hidden-competence-in-human-act">introduced the concept</a> in 2004 to describe intelligent human behavior within complex systems. Their work explored how people can recognize their participation in a larger whole, understand the feedback surrounding them, and act productively even when the system cannot be completely known or controlled.</p><p>That foundation matters. Their work focuses on intelligent human action within complex systems. The challenge I am interested in extending is how we deliberately design the structures, technologies, relationships, incentives, and learning mechanisms through which an institution itself can behave more intelligently as a whole.</p><p>This is where Systems Intelligence meets Strategic Design.</p><p>Systems Intelligence helps people perceive and act intelligently within complexity. Strategic Design helps us deliberately reshape the conditions producing it.</p><p>Together, they form Strategic Design &amp; Systems Intelligence, or SDSI: the applied practice of perceiving, designing, and deliberately shaping complex human, organizational, and technological systems so that institutions can sense, interpret, decide, act, and learn as coherent adaptive wholes.</p><p><strong>Instead of beginning with the product, it begins with the conditions surrounding the product.</strong></p><p><strong>Instead of beginning with the service, it begins with the system producing the service.</strong></p><p><strong>Instead of beginning with the journey, it begins with the forces shaping the journey.</strong></p><p>This shift changes what designers and strategists pay attention to. They still care about human needs, but they also care about organizational constraints. They still care about experience quality, but they also care about incentive structures. They still care about innovation, but they also care about whether the system can absorb and scale what innovation produces.</p><blockquote>Systems Intelligence is not about making everything bigger and more abstract. It is about seeing enough of the whole to make better decisions about the parts.</blockquote><p>That distinction matters because many organizations confuse complexity with sophistication. They create more committees, more governance, more dashboards, more programs, more platforms, and more processes, believing they are building maturity. In reality, they may be creating systems that are harder to understand and slower to adapt.</p><p><strong>A system is not intelligent because it is complex.<br>A system is intelligent when it can sense, learn, respond, and evolve.</strong></p><p>That institutional capacity becomes especially consequential when artificial intelligence begins operating inside the system. I examine that governance challenge in <a href="https://medium.com/@GregLakloufi/artificial-intelligence-needs-systems-intelligence-5128d573bc2c"><em>Artificial Intelligence Needs Systems Intelligence</em>.</a></p><h3>When Optimization Becomes Fragility</h3><p>For much of modern business history, optimization was the dominant ambition. Organizations sought to reduce friction, increase efficiency, improve speed, remove waste, and standardize what worked. These goals still matter. But optimization has a hidden weakness. It often assumes the environment will remain stable enough for the optimized model to keep working.</p><p><strong>In stable environments, optimization creates advantage.<br>In unstable environments, over-optimization creates fragility.</strong></p><p>A supply chain optimized only for cost may collapse when disruption hits. A workforce model optimized only for efficiency may struggle when new skills are suddenly required. A customer service model optimized only for deflection may erode trust when people need human support. A technology architecture optimized only for current needs may become a constraint when the business must pivot.</p><p>Resilience asks a different set of questions. <br>Where does the system have slack? Where can it adapt? Where can it learn? Where are dependencies too concentrated? Where are decisions too slow? Where is knowledge trapped? Where are incentives working against the desired outcome?</p><p>This is not a rejection of efficiency. It is a recognition that efficiency without adaptability can become dangerous.</p><p><strong>The future will not reward organizations that are merely optimized.<br>It will reward organizations that are resilient enough to keep learning.</strong></p><h3>New Skills for a New World</h3><p>For designers, strategists and product leaders, this shift creates both a challenge and an opportunity.</p><p>The challenge is that the work is becoming more complex. It is no longer enough to understand users, map journeys, define personas, facilitate workshops, or design elegant interfaces. Those skills remain valuable, but they are becoming table stakes. <strong>The more valuable skill is the ability to understand the environment in which those artifacts must live.</strong></p><p>The opportunity is that designers are unusually well positioned for this shift, if they are willing to evolve. Good designers already know how to move between ambiguity and clarity. They know how to make invisible dynamics visible. They know how to translate across functions. They know how to hold human needs and business constraints in the same conversation. They know how to prototype possible futures before organizations commit to them.</p><p>Those abilities are not becoming less important.<br>They are becoming more strategic.</p><p>But they need to be aimed at a larger canvas.</p><p>The next generation of design leaders will need to become fluent in systems. They will need to understand how organizations make decisions, how incentives shape behavior, how governance enables or restricts movement, how data flows, how technology architectures influence experience, and how external forces reshape internal priorities.</p><p>This does not mean every designer needs to become an economist, technologist, policy expert, or organizational theorist. It means designers need to become more curious about the forces that shape the work before the work ever reaches the screen, the service blueprint, or the roadmap.</p><p><strong>In practical terms, this begins with asking better questions.</strong></p><p>When a team asks for a new experience, ask what system is producing the current one. When a leader asks for a better product, ask what conditions would allow that product to succeed. When an organization asks for innovation, ask how ideas currently move from concept to reality. When a company asks for AI, ask how intelligence currently flows through the organization and where it gets distorted, delayed, or trapped.</p><p><strong>These questions may feel slower at first. They may even feel inconvenient. But they prevent teams from designing elegant solutions for misunderstood problems.</strong></p><h3>Five Ways to Begin</h3><p>These are not a complete operating model for SDSI. They are practical ways for designers, strategists, and leaders to begin working more intelligently within the systems surrounding their work.</p><p><strong>The first step</strong> is to study the surrounding system before designing the intervention. This means looking beyond the user journey and understanding the broader context that shapes it. Who owns the decision? Who funds the work? Who carries the operational burden? What incentives are driving behavior? What constraints are real, and which ones are inherited assumptions? Where does the current system break, and who absorbs the cost when it does?</p><p><strong>The second step</strong> is to map relationships, not just processes. <br>Traditional process maps show sequence, but systems are often shaped by relationships: between teams, technologies, policies, incentives, data, customers, partners, and external forces. A process map may show how work is supposed to move. A systems map reveals why it behaves the way it does.</p><p><strong>The third step</strong> is to identify leverage points. <br>Not every part of a system matters equally. Some interventions create marginal improvement. Others shift the behavior of the whole. The work is not to fix everything. It is to understand where a small, well-placed change can create disproportionate impact.</p><p><strong>The fourth step</strong> is to design for feedback. <br>Many organizations launch products and services as if implementation is the end of the work. In complex environments, launch is only the beginning. Systems need feedback loops that help teams sense what is changing, learn from real behavior, and adapt before small issues become structural failures.</p><p><strong>The fifth step</strong> is to protect adaptability. <br>As organizations scale, they often remove the very flexibility that made an idea valuable in the first place. Governance, compliance, funding, and delivery processes are necessary, but they can also harden too quickly. The goal is not to eliminate structure. It is to create structures that allow the system to keep learning.</p><p>This is where the work becomes deeply practical. Systems Intelligence is not an abstract intellectual exercise. It changes how teams frame problems, evaluate trade-offs, design operating models, structure decision-making, and measure success. It changes what gets protected as ideas scale. It changes how organizations move from insight to implementation without losing the intent of the work along the way.</p><h3>The Age of Systems</h3><p>Every era asks something different of organizations.</p><p>The Industrial Age rewarded those who could build at scale. The Experience Economy rewarded those who could create meaningful experiences. The Age of Systems will reward those who can learn, adapt, and remain resilient within increasingly complex environments.</p><p>The question is no longer whether organizations can design better products or better services. The question is whether they can design systems capable of adapting to the complexity they have created.</p><p>Products create value.<br>Services deliver value.<br>Systems determine what endures.</p><p><em>(This article draws on publicly available sources and industry research. No proprietary or confidential information has been used. AI tools were used selectively to support research, editing, and visual development; the arguments and perspectives expressed are my own.)</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ec77e50d947f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Designing for Tension: Why Organizations Fail to Translate Ideas into Reality]]></title>
            <link>https://medium.com/@GregLakloufi/designing-for-tension-why-organizations-fail-to-translate-ideas-into-reality-1dc9cc78f959?source=rss-2fa4496fe40b------2</link>
            <guid isPermaLink="false">https://medium.com/p/1dc9cc78f959</guid>
            <category><![CDATA[systems-thinking]]></category>
            <category><![CDATA[innovation]]></category>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[strategic-design]]></category>
            <category><![CDATA[service-design]]></category>
            <dc:creator><![CDATA[Greg Lakloufi]]></dc:creator>
            <pubDate>Sun, 17 May 2026 16:18:03 GMT</pubDate>
            <atom:updated>2026-05-21T21:33:35.118Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*BzDMUpoUBRPWDuIGl1dvag.jpeg" /></figure><h4>How Strategic Design Turns Imagination into Reality Without Losing Its Intent</h4><p>Organizations don’t fail because they lack ideas or execution.<br>They fail because they cannot translate one into the other.</p><p>I recently participated in a three-day design innovation workshop with the US Air Force Pathfinder Program alongside Colonel Jason Trew, PhD (USAF Ret.), former Commandant and Dean of the <a href="https://www.airuniversity.af.edu/SAASS/">School of Advanced Air &amp; Space Studies</a> and author of <a href="https://urldefense.com/v3/__https:/www.airuniversity.af.edu/Portals/10/AUPress/Books/B_0174_Trew_The_Icarus_Solution_1.pdf__;!!Jrgwm_wKFTPpVg!MEPz_hLJqFCHQ4MNhiwwnTHu4LOagkMMd2RyASbE8x-f8Uq4OCQgMFSZOTDxzXkCj1iUqg8WVCR3JB9H9-pdaQ$"><em>The Icarus Solution</em></a>. It was the kind of environment where ideas are quickly confronted by operational reality, shaped by people who understand constraint not as theory, but as a daily condition.</p><p>Over those three days, we moved continuously between two very different modes. At times, conversations were grounded in structure, mission requirements, operational realities, and the discipline required to execute in highly constrained environments. At other moments, the room opened into broader exploration, allowing participants to question assumptions, reframe problems, and imagine possibilities beyond existing systems.<br>What became increasingly clear was not simply that both modes were necessary, but that neither produced meaningful outcomes in isolation. The strongest ideas did not emerge from creativity alone, and the most viable solutions did not emerge from structure alone. They emerged in the tension between the two.</p><p>That friction did not slow the work. It revealed it.</p><p>And once you begin to recognize it, you start to see it everywhere. Including in the way organizations are designed to operate.</p><h3>Control Is Not the Problem. Overcontrol Is.</h3><p>Organizations are built to create order. <br>Strategy defines direction. Processes create consistency. Governance reduces uncertainty. Over time, these mechanisms become the foundation of how work gets done and how success is measured. In stable environments, this architecture enables scale, efficiency, and reliability.</p><p>But most environments are no longer stable.</p><p>For example, a global financial institution recently completed a multi-year transformation of its digital platform. Execution was exemplary. Timelines were met, risks mitigated, dependencies managed. Internally, it was considered a success. Yet within months of launch, adoption lagged significantly behind expectations. Competitors had already redefined what “good” looked like. The organization had delivered precisely what it had planned, but what it had planned no longer mattered.</p><p>This is the paradox. Organizations are becoming more efficient at executing against defined objectives while becoming less effective at adapting those objectives to a changing reality. The system improves its ability to deliver, while losing its ability to sense.</p><p>Control, on its own, is no longer enough.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*X6gtFNIfeCHO44sfCZu1vQ.jpeg" /></figure><h3>Icarus Was Never the Problem</h3><p>That instinct toward control is not new. The story of Icarus has reinforced it for centuries. Do not fly too high. Do not let ambition outrun discipline. Respect limits.</p><p>But this interpretation is incomplete.</p><p>In the original myth, Daedalus, the master craftsman, builds wings for himself and his son Icarus to escape captivity, warning him not to fly too close to the sun. Icarus, driven by ambition, ignores the warning and falls.<br>At a glance, the story feels like a familiar dichotomy. Icarus as ambition and imagination, the instinct to rise. Daedalus as discipline and craft, the structure that makes flight possible. It mirrors the way we often frame creativity and logic, exploration and execution, as opposing forces. As Trew reflected in our conversations:</p><blockquote>“I grew up in an organization that was biased towards Daedalus — special rooms, organizations, and even statues all created in his honor; I didn’t realize that Icarus was not just a fool but perhaps the true hero of the story.”</blockquote><p>As Trew reframes it, Icarus and Daedalus are opposing forces held in necessary tension, interdependent elements of the same system. Icarus represents imagination, ambition, and the instinct to push beyond what exists. Daedalus represents discipline, craft, and the structures that make flight possible. Innovation emerges not by resolving that tension, but by learning how to operate within it.</p><p><strong>Remove Icarus, and nothing new is attempted.<br>Remove Daedalus, and nothing survives.</strong></p><p>This dynamic is not confined to myth. It is embedded in the history of innovation itself. In the early development of airpower, pilots operated ahead of doctrine, experimenting, learning through practice, and shaping the very theories that would later define the field. Practice did not follow strategy. It informed it.</p><p>Yet most modern organizations behave as if ideas must conform before they are allowed to exist. And that misunderstanding does not stay theoretical. It shows up directly in how organizations operate today.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2Lo93AgRrpgzTMmE8mRzQg.jpeg" /><figcaption>The moment Elon Musk challenged legacy aerospace assumptions that had long gone unquestioned.</figcaption></figure><h3>Two Forces. One System. Rarely Designed Together.</h3><p>Organizations do not fail because they lack ideas. <br>They do not fail because they cannot execute.</p><p><strong>They fail because they cannot carry ideas forward without losing what made them valuable in the first place.</strong></p><p>Ideas rarely collapse all at once. More often, they are gradually reshaped by systems optimized for predictability, governance, and scale until the original intent becomes almost unrecognizable.</p><p><strong>This is the failure that rarely appears in reports. Not failure to deliver, but failure to preserve intent.</strong></p><p>To understand why this happens, you have to look at the underlying system. At a fundamental level, every organization operates between two forces.</p><p><strong>Exploration is where ideas are born.</strong> It is fast, ambiguous, and driven by curiosity. It challenges assumptions and expands possibility. Throughout the workshop and the conversations surrounding it, this was often the mode where assumptions loosened, perspectives shifted, and entirely different possibilities began to emerge. It is the space where the system stretches beyond what it currently understands.</p><p><strong>Execution is where ideas survive.</strong> It is structured, disciplined, and oriented toward scale. It translates concepts into systems, processes, and measurable outcomes. It is where ideas are forced to prove their value under constraint.</p><p>Most organizations are designed to optimize one of these forces.<br>Few are designed to integrate them.</p><p>During conversations surrounding the Pathfinder workshop, I had a particularly insightful discussion with Paul Propster, Senior Strategist in the NASA/JPL Office of Strategy, around this exact tension: how organizations balance engineering rigor, operational constraints, and the need to radically rethink entrenched systems. That conversation naturally led to the broader transformation unfolding across aerospace itself. When SpaceX began challenging the economics of spaceflight, Elon Musk pushed engineers, many of them shaped by NASA’s legacy systems, to question assumptions that had long gone unexamined. Why are rockets so expensive to build? Why are they discarded after a single use? Why is reusability treated as an exception rather than a baseline?</p><p>The answers were rarely about physics. They were about precedent. Not because they were optimal, but because they were inherited. Because the system had been designed that way. Because constraints had been accepted rather than interrogated.</p><p>What followed was not incremental innovation. It was a redesign of the system that governed how rockets were conceived, built, and deployed.</p><p><strong>Discipline was not removed. It was restructured.</strong></p><p>Extended development cycles gave way to rapid iteration. Testing became continuous rather than sequential. Failure, within controlled bounds, became a mechanism for learning rather than a point of termination.</p><p><strong>In effect, SpaceX did not choose between exploration and execution. <br>They redesigned how the two interacted.</strong></p><p>The result was not simply lower cost or reusable rockets. It was a fundamentally different relationship between imagination and scale, one where ideas could evolve without being prematurely constrained, and scale without losing their original intent.</p><p>That is the real challenge. Not choosing between these forces, but designing systems where both can operate without canceling each other out. And even when organizations recognize these two forces, something still breaks.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5zLWQSGE4qpFDRC3kuVpSA.jpeg" /></figure><h3>The Gap Where Good Ideas Go to Die</h3><p>The breakdown rarely occurs in exploration or execution themselves. It occurs in the transition between them.</p><p>A financial services firm developed a simplified onboarding experience that significantly improved completion rates during testing. The concept was validated. The value was clear. But as it moved toward implementation, it encountered regulatory constraints, legacy systems, and fragmented ownership across teams. Each adjustment made sense. Together, they altered the experience fundamentally.</p><p>By the time it launched, the solution complied with every requirement and delivered only a fraction of the original value.</p><p>This is the invisible gap.</p><p>It is not a failure of ideas.<br>It is not a failure of execution.</p><p><strong>It is a failure of translation.</strong></p><p>That same dynamic surfaces around the broader challenge of accelerating the design, production, and deployment of drones and unmanned systems within highly constrained environments.</p><p>The question was deceptively simple: how do you move from concept to deployed capability in a fraction of the time, when operating within the realities of government attribution, acquisition protocols, and layers of Pentagon governance designed for control, not speed?</p><p>Traditional acquisition systems optimize for accountability and risk reduction through rigid sequencing and layered governance. The result is often speed sacrificed for certainty, even as the operational environment changes faster than the system can adapt.</p><p>What emerged from our conversations was not a rejection of these constraints, but a different way of thinking about how organizations operate within them. Instead of treating exploration and execution as separate phases, conversations repeatedly returned to the idea that these modes could operate in parallel rather than sequentially. Rapid prototyping alongside evolving requirements. Modular architectures that allowed components to be developed, tested, and deployed independently. Tighter feedback loops between operators, engineers, and decision-makers.</p><p>The shift was subtle but significant:</p><p>From: Design &gt; Approve &gt; Build &gt; Deploy<br><strong>To: Explore &gt; Test &gt; Adapt &gt; Scale</strong></p><p>All happening continuously, not sequentially.</p><p>The challenge was not removing Daedalus, the structures of governance, compliance, and control. Those remain essential in environments where failure carries real consequences.</p><p>The challenge was allowing Icarus, exploration, iteration, and learning, to exist within that system without being suppressed.</p><p>Because in environments where timelines compress and threats evolve rapidly, the inability to translate ideas into deployed capability is not just inefficient.</p><p><strong>It is a strategic vulnerability.</strong></p><h3>The Missing Discipline No One Owns</h3><p>If this gap is so consistent, why don’t organizations have a discipline to address it?</p><p>Execution has been industrialized. Innovation is increasingly structured. But the capability that connects the two remains underdeveloped.</p><p>Translation is not a handoff. It is the discipline of preserving intent while adapting form. It requires identifying what is essential within an idea and ensuring those elements survive as the idea encounters constraints.</p><p>In healthcare, this challenge is particularly visible. A system may design a seamless, patient-centered experience that improves outcomes in pilot environments. But scaling that experience requires integration across clinical workflows, regulatory frameworks, and legacy technologies. Without a disciplined approach to translation, the experience fragments. <strong>What began as transformation becomes incremental improvement.</strong></p><p>This is where most value is lost.</p><p>Not in the idea.<br>Not in the execution.</p><p>In the space between.</p><h3>Design as the Translation Layer</h3><p>This is where design shifts from a discipline to a system function.</p><p>Not as a discipline focused on artifacts, interfaces, or even experiences alone, but as the system responsible for translation.</p><blockquote>At its highest level, design becomes the mechanism that translates imagination into executable reality. It preserves the intent of ideas as they move through strategy, funding, governance, technology, and operations.</blockquote><p>Across the workshop and the conversations surrounding it, this tension became especially visible. The teams that made the most progress were not the most creative or the most structured. They were the ones that could move between both without losing coherence. They understood how to test ideas without diluting them, and how to structure solutions without stripping away their core value.</p><p>This is not a supporting capability. It is an operating necessity.</p><p>In complex systems, the ability to translate is what determines whether organizations evolve or stagnate. But recognizing this role is not enough.</p><h3>Designing for Tension</h3><p>If tension between exploration and execution is inherent, the goal is not to eliminate it. It is to design for it.</p><p>This requires recognizing that these modes operate under different conditions: <br>• Exploration needs space, speed, and tolerance for ambiguity. <br>• Execution requires clarity, alignment, and discipline.</p><p>Forcing them into a single operating model creates friction without progress. More importantly, it requires designing the pathways between them. How ideas are validated, how evidence is generated, how decisions are made, and how ownership transitions without fragmentation.</p><p>These pathways are where most organizations fail. Not because the ideas are weak. But because the system cannot carry them forward.<br>And if this is true, then it changes how performance should be measured.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*GuFicEq8Banif6Kp4E2F3w.jpeg" /></figure><h3>From Insight to Practice: Operating Inside the Tension</h3><p>Traditional metrics emphasize execution: efficiency, predictability, delivery. But in environments defined by change, execution alone is insufficient. A more meaningful measure is the ability to translate. Are insights turning into action? Are ideas becoming durable systems? Is the organization scaling what matters, or simply what is easiest to manage?</p><p>Understanding the tension between exploration and execution is one thing. Operating within it is another. For designers and strategists, this is where the work changes. Not in what you produce, but in how you move.</p><p>Three shifts begin to define the difference:</p><h4><strong>1. Design in Dual Modes, Not Linear Phases</strong></h4><p>Most teams still treat innovation as a sequence. First explore, then define, then deliver. Discovery happens upfront. Execution happens later. Once a direction is chosen, the expectation is to move forward, not revisit.</p><p>In reality, the most effective teams operate differently. <strong>They move back and forth between exploration and structure as insight evolves. Even late in delivery, they continue to test assumptions.</strong> Even early in discovery, they introduce structure to understand what might actually work.</p><p>In practice, this shows up in small but critical moments. A team in delivery pauses to revisit a core assumption after new user feedback. A prototype is tested before requirements are fully locked. A roadmap is treated as directional, not fixed, allowing space to adapt as new information emerges.</p><p>This requires resisting the urge to “lock” ideas too early. <strong>It means allowing concepts to remain fluid long enough to understand what truly matters, while introducing just enough structure to test their viability.</strong> The work is not to separate thinking from doing, but to let each continuously inform the other.</p><p>The question is not “Are we done exploring?”<br>It is “Are we structuring the right thing?”</p><p>But operating in dual modes is only the beginning.</p><h4><strong>2. Protect Intent as Ideas Scale</strong></h4><p>As ideas move through organizations, they are reshaped by constraints. This is inevitable. What is not inevitable is the loss of intent.</p><p>Most organizations attempt to preserve intent through familiar artifacts: a vision statement, a North Star, a set of guiding principles. But too often, these remain static declarations, documents referenced occasionally but rarely used to shape decisions, disconnected from execution.</p><p><strong>High-performing teams do something different. They make intent actionable. They define what must not change, even as everything else adapts. They treat the core value of an idea as something to be actively preserved, not passively assumed.</strong></p><p>In practice, this becomes visible in how decisions are made day to day. Teams anchor every major trade-off against a small set of non-negotiables. When scope is reduced, when timelines compress, when technical constraints emerge, the first question is not “what is easiest to cut?” but “what cannot be compromised?”</p><p>This takes concrete forms. Design principles that travel WITH the idea across teams and are referenced in reviews, not just documented. Clear articulation of the experience or outcome that must survive, even if everything around it changes. Decision frameworks that require teams to justify how a change still aligns with the original intent.</p><p><strong>Without this, every constraint becomes an erosion.<br>With it, constraints become design inputs.</strong></p><p>But even when intent is protected, something still breaks.</p><h4><strong>3. </strong>Design the Pathways, Not Just the End State</h4><p>Most teams focus on either the idea OR the outcome. They invest heavily in defining a vision, or in delivering against a roadmap. Few spend time designing what happens in between. But that is where success is actually determined.</p><p>In practice, this is where things tend to break. An idea leaves a workshop and enters funding conversations. It moves from strategy into delivery teams. It gets translated into requirements, tickets, and timelines. At each step, decisions are made by different functions, often with different incentives. Without intentional pathways, the idea does not evolve coherently. It fragments.</p><p><strong>Designing pathways means making these transitions explicit. It means defining how ideas mature as they move through the system.</strong> What evidence is required to move forward. How trade-offs are evaluated across teams. How ownership shifts without losing continuity.</p><p>You see it in how teams operate. Joint reviews where design, product, engineering, and business stakeholders evaluate decisions together. Clear checkpoints that are based on evidence, not just approval. Shared artifacts that carry intent across pathways instead of resetting it at each transition.</p><p>It also means recognizing that this is not a one-time handoff. It is a continuous process of translation, where ideas are refined without being reduced.</p><p><strong>The strongest teams do not just create solutions.<br>They create the conditions for solutions to survive.</strong></p><p>These shifts do not eliminate tension.<br>They make it usable.</p><h3>The System That Can Fly</h3><p>In the end, all of this points to a different kind of challenge.</p><p>The work is not simply to generate better ideas or execute more efficiently. <strong>It is to design systems that can continuously move between what exists and what could exist without losing either.</strong></p><p>Because the real risk is not ambition. It is building systems that cannot translate it forward.</p><p>The tension between Icarus and Daedalus does not disappear.<br>It intensifies.</p><p>The organizations that endure will not resolve that tension. They will learn to operate within it, continuously. And in doing so, they will build something far more difficult than balance.</p><p>They will build systems that can fly.</p><p><em>(Disclaimer: The companies, projects, and examples mentioned in this article are based on publicly available information and industry best practices. No proprietary or confidential information has been disclosed, and all references are for illustrative purposes only. This article does not represent insider knowledge or violate any non-disclosure agreements. Some AI tools could have been used for basic editing, research and visuals, but the insights and perspectives are entirely human.)</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1dc9cc78f959" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Convergence Era Is Here]]></title>
            <link>https://medium.com/@GregLakloufi/geostrategic-design-the-convergence-era-is-here-eb8768122e83?source=rss-2fa4496fe40b------2</link>
            <guid isPermaLink="false">https://medium.com/p/eb8768122e83</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[strategy]]></category>
            <category><![CDATA[digital-transformation]]></category>
            <category><![CDATA[service-design]]></category>
            <category><![CDATA[future-of-design]]></category>
            <dc:creator><![CDATA[Greg Lakloufi]]></dc:creator>
            <pubDate>Tue, 31 Mar 2026 21:43:39 GMT</pubDate>
            <atom:updated>2026-04-01T13:49:36.287Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rkMlF-Jumvgr8zhSl5NnKg.jpeg" /></figure><h4>We are no longer designing services. We are designing the systems that define reality.</h4><p>A few weeks ago at <a href="https://schedule.sxsw.com/2026/events/PP1148484">SXSW</a>, <a href="https://www.linkedin.com/in/amywebb/">Amy Webb</a>, CEO of Future Today Strategy Group and renowned futurist, introduced the <a href="https://urldefense.com/v3/__https://ftsg.us4.list-manage.com/track/click?u=aa328e1f564f5fd404f866492&amp;id=57d2fc04e9&amp;e=79982d0ae4__;!!Jrgwm_wKFTPpVg!P-mB1YRLd87LoY3kun9MNDVLrwaFzHZ5FRkmZDG4gzUqOhJpNXfqybLWbPpfJrxZomhuKwE_-TS3ge6j$"><em>Convergence Outlook 2026</em></a>, a report that makes one thing impossible to ignore: We are no longer operating in a world of isolated trends, but in one where multiple forces collide and reshape systems at once.</p><p>It does not describe a distant future. It names the reality we are already operating inside, one where change no longer arrives in clean sequences, industries no longer move independently, and strategy can no longer wait for clarity before committing.</p><p>If you have been paying attention, none of this feels surprising. The contours of this shift have been forming for some time. In earlier work, I described it as the emergence of a <a href="https://medium.com/@GregLakloufi/the-rise-of-geostrategic-design-a-compass-for-the-21st-century-2211f1b7211b">Geostrategic</a> layer to Design. What this report does is bring that layer into sharp focus.</p><p>The convergence era is here, and service design is uniquely positioned to evolve and lead in this new era, building on its systemic foundation to address a more complex, interconnected reality.</p><h3>We’ve Entered a New Operating Reality</h3><p>For years, we treated change as something that unfolds in sequence. One trend emerges, another follows, industries adapt, recalibrate, and move forward. That model no longer holds. As Amy Webb puts it, “the future no longer arrives one trend at a time.”</p><p>Change is no longer linear, contained, or predictable in isolation. Multiple forces now collide at once across technology, economics, geopolitics, and human behavior. They don’t move in parallel. They interact, amplify, and destabilize each other.</p><p>Disruption doesn’t just move forward anymore. It spreads sideways, and when it does, the effects are no longer incremental. They are systemic.</p><p>Most organizations can see this happening, but they are still responding with models built for a different era: Strategy assumes time, design assumes boundaries, innovation assumes separation.</p><p>None of those assumptions survive the new concept of convergence. This is why decisions about AI, supply chains, regulation, and infrastructure are now shaping customer experience more than any interface ever could.</p><p>Because convergence is not a cluster of trends. It is a structural shift that cuts across domains, moves non-linearly, redistributes power, and once it takes hold, is difficult to reverse.</p><p>It doesn’t stack on top of existing systems. It rewrites them.</p><p>What once felt improbable becomes inevitable. Value moves, advantage shifts, and boundaries dissolve. And when those boundaries disappear, so does the idea of competition as we understood it.</p><p>You are no longer competing within a category.</p><p>You are competing across ecosystems that did not exist a few years ago.</p><h3>Reality is Already Being Rewritten</h3><p>If this still feels abstract, it shouldn’t. The signals are already here, and they are compounding.</p><p>Identity is no longer something you present. It is continuously inferred, tracked, and verified. Anonymity is quietly disappearing.</p><p>Infrastructure is no longer global. It is fragmenting along geopolitical, energy, and regulatory lines, reshaping how and where systems operate.</p><p>Labor is detaching from value creation as automation scales, concentrating power around ownership of systems rather than human effort.</p><p>At the same time, systems are no longer waiting for users to act. They predict, decide, and intervene before intent is even expressed. Surveillance is no longer exceptional; it is embedded, normalized, and financed by the very services we depend on.</p><p>Emotional support is shifting from human relationships to always-on, machine-mediated environments. And the human body itself is entering the system, as augmentation moves from experimentation to expectation.</p><p>None of these shifts are happening in isolation. Together, they are rewriting something deeper: Behavior, trust, agency, and experience.</p><p>Not gradually, but fundamentally.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uX7Tb5QM18cNXb-EEP5RrA.jpeg" /></figure><h4>The System is Already Making Decisions for You</h4><p>A customer wakes up and checks their phone. Before they take a single action, the system has already acted on their behalf.</p><p>Their device has authenticated them through biometrics. Their news feed has been filtered based on where their data is allowed to flow. Their banking app has flagged a transaction before they notice it. Their health app has adjusted recommendations based on patterns detected overnight.</p><p>They didn’t ask for any of this.<br>No button was pressed. No journey was initiated.</p><p>At the same time, the infrastructure behind these services is shifting. The data they generate is routed through region-specific cloud systems. The AI models interpreting their behavior are tuned differently depending on regulatory constraints. The services they rely on are no longer operating in a global system, but in fragmented, politically defined ones.</p><p>And when they reach out for support, they are just as likely to interact with a machine as a human.</p><p>What looks like a series of separate experiences is not separate at all.</p><p>It is a single system, continuously observing, predicting, and acting across contexts that no longer have clear boundaries.</p><p>And it is already shaping decisions before they are made.</p><h3>The Terrain is Shifting Across Five Domains</h3><p>To understand convergence, you have to see where pressure is building, not as isolated trends but as forces reshaping entire parts of the system at once.</p><ul><li><strong>Compute becomes physical: </strong>Constrained by energy, geography, and geopolitics</li><li><strong>Machines act autonomously</strong>: Decisions and workflows move from humans to systems</li><li><strong>Everything is observed and measured: </strong>Environments and identities become data streams</li><li><strong>Systems begin to feel alive: </strong>AI, sensors, and biology merge into adaptive systems</li><li><strong>Humans shift their dependence</strong>: Care and decision-making increasingly route through machines</li></ul><p>Individually, each of these shifts would be significant. Together, they change the nature of the environment itself.</p><p>This is no longer digital transformation.<br>It is the transformation of reality as a system.</p><h3>The Real Risk is Not Disruption. It’s Hesitation.</h3><p>Most organizations believe disruption is the threat. It isn’t.</p><p>The real risk is seeing what’s coming and failing to act on it. Because that is exactly what most companies will do. They will recognize the signals, analyze the shifts, build the decks, and then wait.</p><p>Traditional planning assumes that clarity comes before commitment. That with enough data, analysis, and time, the right decision will reveal itself. That assumption no longer holds.</p><p><strong>In a convergence environment, waiting is not neutral. It is a decision, and it is usually the wrong one.</strong></p><p>The decisions that matter most now are irreversible. They redefine what your organization becomes, and they must be made before the outcome is clear.</p><p>This is where strategy fundamentally changes. It stops being an exercise in optimization and becomes an act of commitment under uncertainty.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nKCIdbr0W0XIRYmdTzEwcw.jpeg" /><figcaption>NVIDIA CEO Jensen Huang (Chesnot, Getty)</figcaption></figure><h3>Winning Organizations don’t Move Faster. They Decide Differently</h3><p>Not every organization gets caught off guard by shifts like this. But the ones that navigate them successfully don’t just move faster. They think differently.</p><p>They operate with clarity and conviction that most organizations never fully develop. They map where value is moving before it becomes obvious, make explicit bets they cannot easily reverse, choose what they will not do, and define the signals that will trigger action before they need them.</p><p>NVIDIA is a clear example. Long before generative AI became mainstream, it made a decisive shift toward becoming the infrastructure layer for AI. That choice required years of investment, ecosystem building, and a willingness to move beyond its original identity. By the time the market caught up, the position was already established.</p><p>This is no longer just about speed. It is about clarity of position and the willingness to sacrifice what works today to build what will matter tomorrow.</p><h3>This is Where the Service Design Discipline Evolves</h3><p>Not at the edges, but at its core. And once you see it, it becomes difficult to ignore.</p><p>Because the shift described here doesn’t just validate a direction. It exposes a gap between how we have been designing services and what the world now requires us to design.</p><p>In earlier work, I referred to this shift as <a href="https://medium.com/@GregLakloufi/the-rise-of-geostrategic-design-a-compass-for-the-21st-century-2211f1b7211b">geostrategic design</a>. What is now becoming clear is that this is no longer a conceptual extension of the discipline.</p><p>It is becoming its necessary evolution.</p><h3>From Services to Systems to Strategy</h3><p>Service design was never wrong. It was scoped for a different world, one where systems were stable, journeys could be mapped, and touchpoints existed within boundaries that held.</p><p>That world is dissolving.</p><p>Change is no longer happening at the level of the service. It is happening at the level of the system. Technology, economics, geopolitics, and culture are no longer adjacent forces. They are intertwined, shaping each other in real time.</p><p>As they converge, value no longer sits neatly inside a journey or a product. It moves across ecosystems.</p><p>Which means the unit of design must change.</p><p>You are no longer designing services within systems. You are designing the systems themselves. This is the shift from service design to system design to something more expansive. What I previously framed as geostrategic design.</p><p><a href="https://medium.com/@GregLakloufi/the-rise-of-geostrategic-design-a-compass-for-the-21st-century-2211f1b7211b">Geostrategic design</a> is a discipline that operates across forces, environments, and time horizons, where orchestration matters more than optimization and the system is the only place where that orchestration can happen. And once you shift the unit of design from service to system, everything else begins to change.</p><h3>Designing for Systems that Act, not Users who Choose</h3><p>Systems are no longer waiting. They are acting.</p><p>What was once initiated by a user is now triggered by prediction. Decisions are made before intent is expressed, and actions are taken without a clear moment of interaction. Machines don’t just support decisions; they make them, transact, optimize, and operate continuously in the background.</p><p>This is not an extension of service design. It is a redefinition of what the “user” even means.</p><p>In previous work, I described this as a shift toward <a href="https://medium.com/@GregLakloufi/machine-centered-design-is-here-can-you-think-like-a-machine-d5f604ba3c56">machine-centered design</a>. What we are now seeing is that this shift is not isolated. It is part of a broader systemic transformation.</p><p>This fundamentally changes the role of design.</p><p>You are no longer designing for a user moving through a journey. You are designing for systems that interpret, decide, and act on behalf of that user.</p><p>As this shift accelerates, experience itself begins to disappear.</p><p>For decades, we designed experiences as journeys with a beginning and an end. That model is breaking down. <strong>The experience is no longer something you enter. It is something that surrounds you.</strong></p><p>Identity is continuously tracked. Systems act before you do. Emotional support becomes always-on and machine-mediated. There is no clear starting point, no obvious interaction, no defined boundary.</p><p>There is no journey to map.</p><p>The experience becomes ambient, predictive, and invisible. And with that, the work changes from mapping journeys to shaping behaviors, from designing touchpoints to orchestrating continuous environments.</p><p>What used to be visible becomes embedded. What used to be explicit becomes inferred.</p><p>And what used to be designed at the surface now has to be designed at the level of the system.</p><h3>Design is No Longer Neutral</h3><p>As systems become more intelligent, predictive, and embedded, something else begins to surface.</p><p>Power.</p><p>Not as an abstract concept, but as a direct outcome of design decisions.</p><p>When environments are continuously observed, identity is biometric, and behavior is shaped in advance, design stops being about experience alone. It starts determining who has agency and who does not.</p><p>These are not just technological shifts. They are designed systems of influence that shape what people see, feel, trust, and depend on, and ultimately what they believe is normal.</p><p>This changes the responsibility of the discipline.</p><p>As explored in earlier work on geostrategic design, design is no longer confined to experience. It operates within systems shaped by power, governance, and global dynamics.</p><p>You are no longer designing services and experiences. You are designing power structures, behavioral feedback loops, and emotional dependencies at scale.</p><p>Design moves into governance, control, and consequence.</p><p>In the convergence era, it is no longer just an experience discipline.<br>It is a power discipline.</p><h3>Design Moves to the Center of Strategy</h3><p>Service design has long lived downstream, shaping how things work after decisions are made. That separation no longer holds.</p><p>The decisions that matter now are identity-defining. They determine what an organization becomes and how it creates value, and they are inseparable from design.</p><p>Capital allocation, capability building, and structural positioning all shape the systems people experience. These are design decisions.</p><p>Which means design cannot remain a delivery function. It moves upstream into strategy, investment, and transformation, not as support, but as direction.</p><p>Service design stops being about execution.<br>And becomes a way to define the future position of the organization.</p><h3>Designing When There is No Clarity</h3><p>For decades, design relied on understanding the user, reducing uncertainty, and validating before committing. That model assumed time.</p><p>That time is disappearing.</p><p>The decisions that matter now must be made before clarity exists. They reshape identity, lock in direction, and cannot easily be reversed.</p><p>Waiting does not reduce risk.<br>It transfers it.</p><p>Design can no longer rely on validation. It must operate with conviction, not blind conviction, but informed, system-aware conviction that accepts uncertainty as a condition.</p><p><strong>Because the role of Design is no longer to prove what works.<br>It is to decide what to build before the world makes that decision for you.</strong></p><h3>The Shift is no Longer Theoretical</h3><p>If everything here could be reduced to a single idea, it is this:</p><blockquote>“We are moving from designing Services within Systems to designing Systems that define Reality.”</blockquote><p>That is the line.</p><p>And once you see it, it reframes everything.</p><p>This is not an evolution. It is a repositioning from experience to systems, from delivery to direction, from validation to commitment.</p><p>What was previously emerging as a new layer of thinking is now becoming unavoidable. Geostrategic design is not a future concept. It is an emerging necessity, a way to operate across converging forces, design value where it actually moves, and engage with decisions that are deeply consequential and not easily reversible.</p><p>This is not a future to anticipate. It is a reality already taking shape.</p><p>The organizations that recognize this shift early will not just adapt.<br>They will define what comes next.</p><p><em>(Disclaimer: The companies, projects, and examples mentioned in this article are based on publicly available information and industry best practices. No proprietary or confidential information has been disclosed, and all references are for illustrative purposes only. This article does not represent insider knowledge or violate any non-disclosure agreements. Some AI tools could have been used for basic editing, research and visuals, but the insights and perspectives are entirely human.)</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=eb8768122e83" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to Set Up Your Own Agentic Decision Engine]]></title>
            <link>https://medium.com/@GregLakloufi/how-to-set-up-your-own-agentic-decision-engine-c4f71136c1a4?source=rss-2fa4496fe40b------2</link>
            <guid isPermaLink="false">https://medium.com/p/c4f71136c1a4</guid>
            <category><![CDATA[design-strategy]]></category>
            <category><![CDATA[ai-agent]]></category>
            <category><![CDATA[service-design]]></category>
            <category><![CDATA[decision-making]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Greg Lakloufi]]></dc:creator>
            <pubDate>Wed, 18 Mar 2026 22:28:17 GMT</pubDate>
            <atom:updated>2026-03-20T16:15:09.929Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*47AN8X2kFIXHwMnJQ1YWsw.jpeg" /></figure><h4>A practical guide to building an agentic workflow with off-the-shelf AI tools</h4><p>In the previous article, we introduced the <a href="https://medium.com/@GregLakloufi/the-agentic-decision-engine-from-sense-making-to-decision-making-a10f1e20bf7b"><strong>Agentic Decision Engine</strong></a>: A strategy-led workflow where AI agents help designers move from sense-making to decision-making. But understanding the concept is only the first step.</p><p>The real question is: H<strong>ow do you actually set one up?</strong></p><p>This is a very practical guide for designers and strategists who want to move from theory to execution using available tools like ChatGPT or Claude. You can do this as a single designer, on your own, without asking for permission <em>(Check your company data policies before sharing proprietary data with an open AI agent)</em>.</p><p>You do not need to start with complex systems or engineering teams.<br>You start small. You structure the workflow. And over time, you build not just one agent, but a <strong>system of agents working together to support decisions across your organization</strong>.</p><h3>1. Start with the Decision</h3><p>Do not begin with AI.<br>Begin with one decision that matters. If you are unsure where to start, pick a decision you are already involved in this week. It does not need to be perfect. It just needs to be real.</p><p>Ask:<br>• What decision are we trying to improve?<br>• Where are we operating with incomplete context?<br>• Where are we relying too heavily on intuition or fragmented data?</p><p>Write it clearly:<br><em>We need to improve the quality of [decision] by increasing the context available before it is made.</em></p><p>Then define:<br>• The business objective<br>• The expected outcome<br>• The key trade-offs</p><p>Also identify:<br>• Who makes the decision<br>• Who influences it<br>• Who executes it</p><p>If there is no clear decision owner, stop. <br>You do not yet have a real use case.</p><p>Finally, map how the decision is made today:<br>• Where are the gaps?<br>• Where is context missing?<br>• Where is the process manual or siloed?</p><p>That is where your agentic system will create value.</p><h3>2. Define the Context the System Needs</h3><p>Once the decision is clear, define the signals behind it.<br>Think in two layers.</p><h4>• Core signals</h4><p>The minimum data required to make the decision:<br>• Sales<br>• Inventory<br>• Pricing<br>• Operational data</p><h4>• Expanded contextual signals</h4><p>The signals that improve judgment:<br>• Weather<br>• Events<br>• Competitor activity<br>• Regional trends<br>• Behavioral patterns</p><p>Then map where these signals and data come from <em>(see section 4 for your potential data sources)</em>:<br>• Internal systems<br>• External data sources<br>• Manual inputs</p><p>You are not building yet.<br>You are defining <strong>what your system needs to see</strong>.</p><h3>3. Design the Agent System</h3><p>Do not build one complex agent.<br>Start with simple roles: <em>(see the complete classes of AI Agents in our previous article </em><a href="https://medium.com/@GregLakloufi/the-agentic-decision-engine-from-sense-making-to-decision-making-a10f1e20bf7b"><strong><em>Agentic Decision Engine</em></strong></a><strong><em>)</em></strong></p><h4>• Signal Agent</h4><p>Collects and organizes inputs, surfacing what changed.</p><h4>• Context Agent</h4><p>Interprets patterns, identifying what matters.</p><h4>• Recommendation Agent</h4><p>Proposes actions, explaining trade-offs.</p><p>Optionally later:</p><h4>• Monitoring Agent</h4><p>Tracks outcomes, improving the system over time.<br>Treat each agent like a team member.</p><p>Define:<br>• Purpose<br>• Inputs<br>• Outputs<br>• Boundaries</p><p>Then design the output before anything else.</p><p>Ask: <strong>What should the decision-maker receive?</strong></p><p>Here is a useful structure to start with:<br>• What changed<br>• Why it matters<br>• Options<br>• Recommended action<br>• Confidence level<br>• Risks or unknowns</p><p>If the output is not useful, the system will not be adopted.</p><p>What you are designing is not a single tool.<br>It is the foundation of an <strong>agentic system</strong> that can scale from one decision to many, from one designer to an entire team.</p><h3>4. Configure Your First Workflow</h3><p>This is where most teams get stuck. It is also where things become surprisingly simple.</p><p>You can set up your first version in ChatGPT, Claude, or similar tools in less than an hour.</p><p>If you can use ChatGPT, you can build your first agent system.</p><p>No engineering required.</p><h4>• Step 1: Create your first agent (inside GPT or Claude)</h4><p>Open your tool and create a <strong>new chat or custom GPT/project</strong>.</p><p>Give it a name:<br>Signal Agent, Context Agent, Decision Assistant…</p><p>Then paste a simple instruction like this:</p><p><em>You are a decision-support agent.<br>Your role is to analyze business signals and produce a short decision brief.</em></p><p><em>You do not make final decisions.<br>You expand the context before a decision is made.</em></p><p><em>Always return:<br>what changed<br>why it matters<br>risks<br>recommended actions<br>confidence level</em></p><p>Congratulations! You now have your first working AI agent!<br>It will not be perfect. That is expected. The goal is not to get it right on the first try, but to start improving how you structure decisions.</p><h4>Note: Where Does the Data Come From?</h4><p>At this stage, you do not need perfect or fully integrated data systems.<br>Start with what you already have access to.</p><p>Most designers can begin with a mix of:<br><strong>Internal data:<br></strong>• Dashboards (Tableau, Power BI, Looker)<br>• Spreadsheets (Excel, Google Sheets)<br>• CRM tools (Salesforce, HubSpot)<br>• Product analytics (Amplitude, Mixpanel)<br>• Internal reports or slide decks</p><p><strong>External signals:<br></strong>• Weather data (simple forecasts from public sources)<br>• Market trends (Google Trends, industry reports)<br>• Competitor activity (websites, pricing pages, newsletters)<br>• News and events (Google News, RSS feeds, social platforms)</p><p><strong>Manual inputs: <br></strong>• Stakeholder insights<br>• Field notes<br>• Customer feedback<br>• Operational updates from teams</p><p>At the beginning, it is completely acceptable to copy and paste data manually into your agent.</p><p>The goal is not automation.<br><strong>The goal is to structure the decision and expand the context around it</strong>.</p><p>As the workflow proves valuable, data access can be automated later with engineering support.<br>If you can access it, you can use it. You do not need a perfect data pipeline to get started.</p><h4>• Step 2: Feed it structured input (not random data)</h4><p>Do not just paste raw data from the sources mentioned above.</p><p>Create a simple agent input block like this and reuse it every time:</p><p><em>Decision:<br>Weekly inventory allocation</em></p><p><em>Objective:<br>Improve availability while reducing excess inventory</em></p><p><em>Data:<br>Sales: [paste data]<br>Inventory: [paste data]<br>External signals: [paste weather, events, etc.]<br>Constraints: [paste issues or limitations]</em></p><p>Paste this into your agent. You now have a <strong>repeatable workflow</strong>.</p><h4>• Step 3: Improve the output format</h4><p>If the output feels messy, tighten it.</p><p>Add this to your prompt:<br><em>Keep the response under 200 words.<br>Be concise. Prioritize clarity over completeness.</em></p><p>This alone dramatically improves usability.</p><h4>• Step 4: Create a second agent (Context Agent)</h4><p>Open a <strong>new chat</strong> (or new GPT).</p><p>Paste:<br><em>You analyze decision summaries and identify what matters most.</em></p><p><em>Focus on:<br>key patterns<br>trade-offs<br>risks<br>uncertainty</em></p><p><em>Return:<br>top insights<br>key trade-offs<br>areas of uncertainty</em></p><p>Now copy the output from your first agent into this one.</p><p>You’ve just created a <strong>2-step agent system</strong>.</p><h4>• Step 5: Add a third agent (Recommendation Agent)</h4><p>Create another new chat.</p><p>Paste:<br><em>You generate decision recommendations based on analysis.</em></p><p><em>Provide 2–3 options.<br>Explain trade-offs clearly.</em></p><p><em>Return:<br>recommended actions<br>why<br>risks<br>confidence</em></p><p>Now feed it the Context Agent output.</p><h4>• Step 6: Run the full workflow manually</h4><p>Your workflow is now:</p><ol><li>Paste data → Signal Agent</li><li>Pass output → Context Agent</li><li>Pass output → Recommendation Agent</li><li>Review with human</li></ol><p>No automation needed yet.</p><p>This is your <strong>Agentic Decision Engine v1</strong>.</p><h4>• Step 7: Add simple guardrails</h4><p>Update each agent with:</p><p><em>If data is missing, say so.<br>Do not invent information.<br>Clearly label uncertainty.</em></p><p>This is what makes the system trustworthy.</p><h4>• Step 8: Save and reuse</h4><ul><li>Save prompts in a doc</li><li>Save your input template</li><li>Reuse weekly or daily</li></ul><p>Consistency is what makes this work.</p><h4>• Step 9: Only automate once it works</h4><p>Once this manual flow is useful:</p><p>Then connect to:<br>• Google Sheets<br>• Slack / Teams<br>• Zapier / Make</p><p>Do not automate before it works manually. <br>Once the workflow is working, the next step is to operationalize it.</p><p>You can now create as many agents as you need, based on your current and future needs. Imagine a team of AI agents working tirelessly and seamlessly to support and guide your decision-making.</p><p>From here, each new agent you add expands your system. Over time, this evolves into an orchestrated network of agents supporting multiple decisions across the organization. In effect, you are building an <strong>agentic layer around your decision-making</strong>, not just a single workflow.</p><h3>5. Pilot, Measure, and Scale</h3><p>Start small.</p><p>Choose:<br>• One decision<br>• One team<br>• One use case</p><p>Run a short pilot.</p><p>Keep a human in the loop:<br><strong>Agent output → human review → decision</strong></p><p>Measure:<br>• Time saved<br>• Quality of insights<br>• Decision confidence<br>• Business outcomes</p><p>Log everything:<br>• Recommendations<br>• Decisions<br>• Outcomes</p><p>Then improve:<br>• Remove noise<br>• Add missing signals<br>• Refine outputs</p><p>Only then should you scale. <br>This is where the shift happens: from a single workflow to a repeatable decision capability that teams can adopt and extend.</p><p>Expand gradually:<br>• More signals<br>• More users<br>• More decisions<br>• More automation</p><h3>A Simple Way to Start This Week</h3><p>If you want to begin immediately:</p><p><strong>Day 1</strong><br>Pick one decision and define the objective.<br>For example:<br>• Which users should we prioritize this week?<br>• Which feature should we improve next?<br>• Which customer segment needs attention?</p><p><strong>Day 2</strong><br>Create one agent and one input template</p><p><strong>Day 3</strong><br>Add a second agent and chain the outputs</p><p><strong>Day 4</strong><br>Add a recommendation agent and test scenarios</p><p><strong>Day 5</strong><br>Review with a stakeholder and refine</p><p>That is more than enough to get you started.</p><h3>Designing How Decisions Are Made</h3><p>The goal is not to build AI agents.<br>It is to improve how decisions are made.</p><p>Start with one decision.<br>Expand the context around it.<br>Keep a human in the loop.<br>Then scale.</p><p>Over time, what begins as a simple workflow becomes something much more powerful: <strong>A system of agents supporting decisions across teams, functions, and domains.</strong></p><p>This is how designers move upstream.</p><p>From shaping experiences, to shaping services, to shaping how organizations think, decide, and act.</p><p>That is how an Agentic Decision Engine becomes real.</p><p><em>(Disclaimer: The companies, projects, and examples mentioned in this article are based on publicly available information and industry best practices. No proprietary or confidential information has been disclosed, and all references are for illustrative purposes only. This article does not represent insider knowledge or violate any non-disclosure agreements. Some AI tools could have been used for basic editing, research and visuals, but the insights and perspectives are entirely human.)</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c4f71136c1a4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Agentic Decision Engine: From Sense-Making to Decision-Making]]></title>
            <link>https://medium.com/@GregLakloufi/the-agentic-decision-engine-from-sense-making-to-decision-making-a10f1e20bf7b?source=rss-2fa4496fe40b------2</link>
            <guid isPermaLink="false">https://medium.com/p/a10f1e20bf7b</guid>
            <category><![CDATA[design-strategy]]></category>
            <category><![CDATA[service-design]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[ai-agent]]></category>
            <category><![CDATA[decision-making]]></category>
            <dc:creator><![CDATA[Greg Lakloufi]]></dc:creator>
            <pubDate>Tue, 17 Mar 2026 00:24:40 GMT</pubDate>
            <atom:updated>2026-09-16T19:49:59.643Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-8hanfm5jHWPe7XeJmKSFQ.jpeg" /></figure><h4>How Agentic AI Helps Designers Turn Strategy into Context-Driven Decisions</h4><p>Most conversations today about AI are dominated by one obsession: Speed.</p><p>Faster research. Faster insights. Faster execution.</p><p>But speed is rarely the only bottleneck in complex strategic decision-making. A deeper limitation is often context: Incomplete information, fragmented signals, and narrow perspectives.</p><p>Most organizations are trying to move faster, but the real advantage comes from seeing more clearly.</p><p>The true power of AI is not acceleration.<br>It is context amplification.</p><blockquote>“AI will not replace human judgment. It will redefine the amount of context available before <strong>a decision is made</strong>.”</blockquote><p>For the first time, organizations can generate, synthesize, and evaluate vast landscapes of information before making decisions. When used correctly, AI does not simply automate work. It expands the contextual field in which judgment occurs.</p><p>This article explores how agentic AI can support the journey from sense-making to decision-making, helping designers and strategists gather information, evaluate signals, generate recommendations, and justify strategic choices.</p><p>More importantly, it introduces a simple idea: AI should not be treated just as a productivity engine. It should be designed as a context engine built to enable strategy.</p><p>Strategy defines the direction &gt; Designers frame the questions &gt; Agentic AI gathers and interprets the signals needed to execute that strategy with greater clarity.</p><h3>From Speed to Context</h3><p>The early narrative around AI has mostly focused on productivity gains. Generate faster. Produce more. Automate work.</p><blockquote>“Strategic work does not fail because it is slow. It fails because it lacks context. Speed without context simply drives organizations faster in the wrong direction.”</blockquote><p>Poor business decisions usually emerge from:<br>• Incomplete information<br>• Fragmented signals<br>• Biased interpretation<br>• Weak connections between insights and strategic objectives</p><p>Today, AI changes this dynamic.</p><p>By generating and evaluating vast volumes of information, agentic systems allow designers to construct richer contextual landscapes before making decisions.</p><p><strong>The result is not simply faster work.</strong><br> <strong>It is a richer foundation for judgment.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*xK3wxLtA33nziH8idX8nzw.jpeg" /></figure><h4>Example: Retail Intelligence at Walmart</h4><p>Walmart increasingly uses AI systems that monitor supply chains, consumer demand patterns, weather forecasts, and logistics signals simultaneously. Instead of reacting to individual data points, these systems build a contextual model of demand across thousands of stores. When weather patterns predict increased demand for certain products in specific regions, supply adjustments happen automatically.<br>The decision is not faster simply because it is automated. It is better because it is informed by a far broader context.</p><h3><strong>The Agentic Decision Engine</strong></h3><p>To understand how AI supports strategic work, it helps to think in terms of what we call the<strong> Agentic Decision Engine (ADE)</strong>.</p><p>This engine combines AI agents, data intelligence, and strategic evaluation to help designers operationalize strategy and move confidently from sense-making to decision-making.</p><p>The strategy itself does not come from the agentic engine. It comes from human judgment, business objectives, and design leadership.</p><p>Rather than replacing human judgment, it expands the contextual intelligence available before a decision is made.</p><blockquote>“The future of AI is not faster answers. It is deeper situational awareness.”</blockquote><p>Instead of simply accelerating tasks, agentic AI expands the informational landscape that surrounds a decision.</p><p>The agentic decision engine unfolds in four stages:<br>1. Gathering information<br>2. Evaluating and contextualizing signals<br>3. Generating strategic recommendations<br>4. Justifying and validating decisions</p><p>Each stage increases the depth and clarity of context, allowing organizations to move from fragmented signals toward confident action.</p><p>In this sense, AI becomes less of an automation tool and more of a strategic intelligence layer embedded inside the organization.</p><p>Agentic AI can support a structured decision flow composed of four core stages:</p><h3>1. Gathering Information</h3><p>Once strategic objectives are defined, the first role of AI agents is expanding access to the data required to pursue them. This includes:<br>• Market intelligence<br>• Customer signals<br>• Operational data<br>• Competitive analysis<br>• Cultural and geopolitical context</p><p>At this stage, the goal is purely information breadth.<br>AI agents collect signals across multiple sources and platforms, helping designers move beyond limited datasets.</p><p>We will explore the seven classes of AI agents later, but two types are particularly relevant at this stage:</p><ul><li><strong>Simple Reflex Agents<br></strong>Agents that respond to immediate inputs using predefined rules.</li><li><strong>Model-Based Agents<br></strong>Agents that build an internal model of the environment to interpret new information more intelligently.</li></ul><p>Together, these agents help construct the initial knowledge landscape.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TTAae5tsG-Y3k15YcGqI8g.jpeg" /></figure><h4>Example: Spotify’s Music Discovery Intelligence</h4><p>Spotify uses AI systems that constantly gather signals from listening behavior, playlist interactions, skip rates, and global listening patterns.<br><strong>Simple agents</strong> react immediately to listening signals while <strong>model-based agents</strong> build complex user taste profiles.<br>This continuous information gathering allows Spotify to understand evolving music preferences at a global scale. The result is not simply music recommendations. It is an evolving map of cultural taste.</p><h3>2. Evaluating and Contextualizing Information</h3><p>Once information is gathered, the next challenge is interpretation.</p><p>Raw data alone does not create insight. Strategic intent determines which signals matter and which can be ignored. AI agents can help evaluate patterns, detect anomalies, and organize information into meaningful structures.</p><p>Here two additional types of agents become critical:</p><ul><li><strong>Goal-Based Agents<br></strong>These agents evaluate information based on defined objectives. They help determine which signals matter and which do not.</li><li><strong>Utility-Based Agents<br></strong>These agents compare potential outcomes and prioritize the options that maximize value.</li></ul><p>In strategic design, this stage transforms scattered signals into coherent context. This is where sense-making happens.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*t9fwHQDWeIacfVEV1DH_KA.jpeg" /></figure><h4>Example: Healthcare Diagnostics at Mayo Clinic</h4><p>The Mayo Clinic uses AI-assisted diagnostic systems that analyze patient data, imaging, research literature, and historical medical records.<br><strong>Goal-based agents</strong> evaluate possible diagnoses relative to patient symptoms. <strong>Utility-based agents</strong> help prioritize treatment paths based on outcomes and risks.<br>Doctors still make the final decisions. But AI dramatically improves the contextual landscape in which those decisions are made.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qDY_ElWLiLTJkNLcThgacA.jpeg" /></figure><h4>Example: Central Banks Interpreting Economic Signals</h4><p>Central banks have traditionally relied on economic reports that appear weeks or months after economic activity occurs.<br>Today, institutions such as the Bank of England and the Federal Reserve are experimenting with AI systems that analyze real-time economic signals.<br>These systems ingest massive streams of information, including shipping data, satellite imagery of industrial activity, retail transactions, job postings, and financial market signals.<br><strong>Goal-based agents</strong> evaluate how these signals relate to macroeconomic targets such as inflation, employment, and productivity. <strong>Utility-based agents</strong> help policymakers explore possible policy responses and their expected outcomes.<br>Human policymakers still make the final decisions. But AI dramatically expands the contextual picture in which those decisions occur.<br>Instead of relying only on lagging historical indicators, economic leaders can begin interpreting emerging economic patterns as they unfold.</p><h3>3. Preparing Strategic Recommendations</h3><p>Once context is clear, AI can assist in building structured recommendations.</p><p>This is where agentic systems begin generating potential pathways forward, including:<br>• Service innovations<br>• Operational improvements<br>• Market expansion strategies<br>• Ecosystem partnerships<br>• Technology integration opportunities</p><p>At this stage, AI agents help generate strategic options while designers frame the problem, evaluate trade-offs, and determine which directions align with the organization’s strategy.</p><p>The key business value often emerges through three capabilities:<br><strong>• Automating repetitive tasks: </strong>Reducing operational friction and freeing human time for strategic thinking.<br><strong>• Learning and adapting: </strong>Learning agents refine recommendations as new information appears.<br><strong>• Working across platforms: </strong>Agentic systems synthesize data across tools, departments, and systems.</p><p>This creates recommendations that are more informed and more adaptive.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DKPxoc0zOUmJyFZjEUtbRQ.jpeg" /></figure><h4>Example: Automotive Design at BMW</h4><p>BMW uses AI-driven simulation environments to explore thousands of design variations and manufacturing options before a vehicle reaches production.<br>AI systems analyze aerodynamic performance, material constraints, cost models, and sustainability targets. Design teams review the generated options and select the most promising directions.<br>Instead of replacing designers, AI expands the space of possibilities designers can explore.</p><h3>4. Justifying the Decision</h3><p>A strong recommendation must ultimately lead to a decision.</p><p>AI can support decision justification by mapping proposals to strategic goals, constraints, and expected outcomes.</p><p>This stage focuses on answering critical questions:<br>• What problem are we solving?<br>• What strategic objective does this support?<br>• What value does this create?<br>• What risks remain?</p><p>By structuring these relationships, AI agents help make decisions more transparent and defensible.</p><p>The role of AI here is not to replace judgment but to strengthen the logic behind it. The final decision remains a human responsibility tied to strategic accountability.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kxdG1Qqc09e4DDwxCtorgA.jpeg" /></figure><h4>Example: Netflix Content Investment</h4><p>Netflix relies heavily on data intelligence when deciding which original content to fund.<br>AI models analyze viewing patterns, genre preferences, regional demand, actor popularity, and storytelling trends. These insights do not dictate the decision. Instead they help justify investments by linking creative ideas to measurable audience signals. Human creative leadership remains central, but the strategic rationale becomes clearer.<br>These capabilities rely on several foundational agent architectures that designers should understand at a conceptual level.</p><p><strong>NOTE:</strong> If you’re wondering how to actually build and deploy this in practice, we’ve put together a simple <a href="https://medium.com/p/c4f71136c1a4?postPublishedType=initial"><strong>step-by-step guide</strong></a> on setting up your own Agentic Decision Engine.</p><h3>Seven Agent Architectures Designers Should Understand</h3><p>There is no single taxonomy for modern AI agents, but several foundational agent architectures provide useful mental models for designers. These concepts help explain how agents can react, model environments, pursue goals, evaluate tradeoffs, learn, coordinate, and operate across levels of abstraction:</p><ul><li><strong>Simple Reflex Agents<br></strong>Operate through predefined rules and immediate responses. These agents react directly to specific inputs without considering past events or broader context. If a condition is met, a corresponding action is triggered, much like a basic automation rule responding instantly to a signal.</li><li><strong>Model-Based Agents</strong><br>Maintain an internal representation of their environment to make more informed decisions. Instead of reacting only to immediate inputs, these agents track how conditions evolve over time and use that internal model to interpret new information within a broader context.</li><li><strong>Goal-Based Agents</strong><br>Evaluate possible actions based on clearly defined objectives. Rather than simply reacting to conditions, these agents consider different paths forward and choose the one most likely to achieve a specific goal, such as improving efficiency, reducing risk, or reaching a target outcome.</li><li><strong>Utility-Based Agents</strong><br>Assess multiple possible outcomes and select the option that delivers the greatest overall value. These agents weigh trade-offs between competing factors such as cost, speed, performance, or risk in order to determine which decision produces the best overall result.</li><li><strong>Learning Agents</strong><br>Continuously improve their performance by analyzing feedback and new data. Over time, these agents adapt their behavior, refine predictions, and update their models, allowing them to respond more effectively as environments and conditions change.</li><li><strong>Hybrid Agents</strong><br>Hybrid agents combine multiple reasoning mechanisms within a single system. For example, an agent may include simple reflex responses for fast reactions, a model-based component to understand context, and a learning module that improves performance over time. The goal of a hybrid architecture is to balance speed, adaptability, and strategic evaluation within a single decision system. In practice, many modern AI systems are hybrid because complex environments require more than one type of reasoning.</li><li><strong>Hierarchical Agents<br></strong>Hierarchical agents organize decision-making across multiple levels of control. Higher-level agents define goals or strategic direction, while lower-level agents execute specific tasks required to achieve those goals. Each layer focuses on a different level of abstraction, allowing complex problems to be broken down into manageable decisions. For example, a strategic agent may decide to optimize supply chain resilience, while subordinate agents manage routing, inventory adjustments, and logistics execution.</li></ul><p>Understanding these models helps design strategists build better AI-assisted workflows.</p><h3>Linking Agentic AI to Strategic Objectives</h3><p>Technology alone does not create value.<br>Value appears when technology is aligned with strategy.</p><blockquote>“Strategy defines the destination. Agentic AI helps organizations navigate the terrain.”</blockquote><p>Before deploying AI systems, organizations should answer several strategic questions:<br>• What are the biggest opportunities we face?<br>• What obstacles slow down our decisions?<br>• Where could efficiency unlock scale?<br>• Which strategic initiatives require stronger intelligence?</p><p>Once these questions are clear, AI can support the organization in:<br>• Unlocking operational bottlenecks<br>• Accelerating strategic initiatives<br>• Scaling intelligence across teams<br>• Improving decision quality</p><p>In this sense, agentic AI becomes a form of organizational cognition. It extends how institutions perceive, interpret, and act on complex environments.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*eGg-4oTemO7qcS6aL_M5Og.jpeg" /></figure><h4>Example: Amazon Logistics Optimization</h4><p>Amazon uses AI systems to orchestrate global logistics across warehouses, transportation networks, and delivery routes. Instead of optimizing isolated operations, agentic systems continuously balance cost, speed, capacity, and demand across the network. The result is a logistics ecosystem capable of adapting dynamically to changing conditions.<br>AI becomes a strategic enabler rather than simply a tool.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*oxTda8m97eTbr93dTqfDXw.jpeg" /></figure><h4>Example: Geopolitical Risk Monitoring in Global Supply Chains</h4><p>Global companies now operate in an environment shaped by geopolitical instability, trade fragmentation, and regulatory shifts.<br>Logistics companies such as Maersk and DHL deploy AI-driven intelligence platforms that monitor thousands of signals across the global system, including shipping flows, political developments, sanctions regimes, energy markets, and regional conflict indicators.<br>These systems continuously analyze news sources, trade data, port congestion, and diplomatic announcements. When signals suggest potential disruption, companies can simulate alternative supply chain routes and sourcing strategies.<br>Instead of reacting after disruption occurs, organizations gain the ability to anticipate geopolitical friction and reposition supply networks in advance.<br>In this environment, AI becomes a strategic sense-making layer for navigating an increasingly complex global economy.</p><h3>Strategy First, Technology Second</h3><p>The biggest mistake organizations make with AI is starting with the technology. They experiment with tools before defining strategic intent.</p><p>The correct order is the opposite.<br>First define:<br>• The strategic objective<br>• The expected outcomes<br>• The decision challenges<br>• The business value</p><p>Only then should technology be deployed to support those goals.<br>AI should amplify strategy, not define it. But what does this look like in practice?</p><h3>From Strategy to Execution: A Simple Example</h3><p>To make this concrete, consider a fictional retail brand facing declining in-store performance.</p><p>A service designer is tasked with improving demand forecasting and reducing stockouts across regions. The team defines a clear objective: Improve product availability while reducing excess inventory.</p><p>From there, the designer frames the problem:<br>• What signals are we missing?<br>• Where are decisions being made with incomplete context?<br>• Which teams are operating in silos?</p><p>AI agents are deployed to gather and synthesize signals across the system:<br>• Real-time sales data<br>• Regional weather patterns<br>• Local events and demand spikes<br>• Supply chain constraints<br>• Competitor pricing signals</p><p><strong>Model-based</strong> and <strong>learning agents</strong> build a dynamic view of demand across regions.</p><p><strong>Goal-based</strong> and <strong>utility-based agents</strong> evaluate trade-offs between cost, availability, and speed.</p><p>The system begins generating recommendations for inventory allocation and replenishment strategies.</p><p>At this stage, the designer and business stakeholders review the outputs, align them with strategic priorities, and validate the direction.</p><p>Finally, the work is operationalized.</p><p>The designer collaborates with engineering teams to translate these capabilities into production systems:<br>• Defining data pipelines<br>• Specifying agent behaviors<br>• Integrating with existing platforms<br>• Establishing feedback loops for continuous learning</p><p>The result is not just a model or a dashboard. It is a decision system embedded into the organization, continuously improving how decisions are made.</p><h3>Where to Start with Agentic AI</h3><p>For many designers and strategists, the question is not <strong><em>what agentic AI is</em>,</strong> but <strong><em>how to actually begin</em>.</strong></p><p>The starting point is simpler than it appears.</p><p>You do not begin by building complex AI systems.<br>You begin by identifying <strong>one decision worth improving</strong>.</p><h4>• Step 1: Identify a High-Value Decision</h4><p>Focus on a decision that is:<br>• Repeated frequently<br>• Currently based on incomplete information<br>• Impactful to business performance</p><p>For example: Inventory allocation, pricing adjustments, or customer segmentation.</p><h4>• Step 2: Map the Signals Behind the Decision</h4><p>Before thinking about AI, map the inputs:<br>• What data informs this decision today?<br>• What signals are missing?<br>• Where are the blind spots?</p><p>This defines what your future agents will need to observe.</p><h4>• Step 3: Start with Off-the-Shelf Agents</h4><p>You do not need to build everything from scratch. Most organizations begin by using existing tools:<br>• LLM-based agents <em>(OpenAI, Anthropic, etc.)</em><br>• Workflow automation platforms <em>(Zapier, Make, etc.)</em><br>• Data platforms <em>(Snowflake, Databricks)</em><br>• BI tools with AI layers</p><p><strong>NOTE:</strong> See our simple <a href="https://medium.com/p/c4f71136c1a4?postPublishedType=initial"><strong>step-by-step guide</strong></a> on setting up your own Agentic Decision Engine.</p><p>At this stage, agents can be simple:<br>• Monitoring signals<br>• Summarizing data<br>• Generating recommendations</p><p>Think of this as <strong>augmenting the decision</strong>, not automating it.</p><h4>• Step 4: Define Agent Roles (Not Just Tools)</h4><p>Instead of thinking in terms of tools, think in terms of roles:<br>• A <strong>signal agent</strong> gathers and aggregates data<br>• A <strong>context agent</strong> interprets patterns<br>• A <strong>decision-support agent</strong> generates recommendations</p><p>This aligns directly with your Agentic Decision Engine.</p><h4>• Step 5: Collaborate with Engineering to Scale</h4><p>Once the value is proven, the system can be industrialized.</p><p>This is where designers partner with engineering to:<br>• Formalize data pipelines<br>• Define agent behaviors and logic<br>• Integrate with operational systems<br>• Establish monitoring and feedback loops</p><p>At this stage, the system evolves from a prototype into a <strong>decision infrastructure</strong>.</p><h4>• Step 6: Decide When to Build vs Buy</h4><p>A simple rule:<br><strong>• Buy</strong> when the capability is common (summarization, prediction, automation)<br><strong>• Build</strong> when the capability is core to your competitive advantage</p><p>Most organizations start by buying and progressively build where differentiation matters.</p><h4>Final Thought</h4><p>The goal is not to build AI agents for their own sake. It is to <strong>improve how decisions are made</strong>.</p><p>Start small. Focus on one decision. Expand from there.</p><p>The organizations that win will not be the ones that adopt AI the fastest, but the ones that redesign how decisions are made.</p><h3>The Age of Contextual Intelligence</h3><p>Agentic AI represents a profound shift for designers and strategists.<br>Not because it makes work faster. But because it expands the contextual intelligence available before decisions are made.</p><p>By gathering vast landscapes of information, interpreting patterns across systems, generating strategic options, and supporting decision justification, AI strengthens the entire chain between insight and action.</p><p>This shift carries an important implication for designers.</p><p>As AI becomes embedded in decision environments, designers are no longer only shaping experiences or services. Designers are increasingly <strong>shaping systems of intelligence</strong> that help organizations interpret complexity and execute strategy with greater precision. In many organizations, designers are quietly becoming architects of decision environments, structuring how intelligence flows from data to action.</p><p>The role of design begins to move upstream.<br>From designing interactions to designing services, to designing decision ecosystems. But more context does not automatically produce better decisions. Organizations must also determine which signals matter, who has authority to act, when human judgment is required, how interventions are governed, and how outcomes feed back into future decisions.</p><p>In a world defined by geopolitical turbulence, economic volatility, and accelerating technological change, the organizations that succeed will not simply move faster. They will see more clearly.</p><p>And in that future, the most powerful use of AI will not be automation.<br>It will be deeper contextual awareness.</p><p><em>(Disclaimer: The companies, projects, and examples mentioned in this article are based on publicly available information and industry best practices. No proprietary or confidential information has been disclosed, and all references are for illustrative purposes only. This article does not represent insider knowledge or violate any non-disclosure agreements. Some AI tools could have been used for basic editing, research and visuals, but the insights and perspectives are entirely human.)</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a10f1e20bf7b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The New Right Stuff: Mastering Failure Points and Unhappy Paths]]></title>
            <link>https://medium.com/@GregLakloufi/the-new-right-stuff-mastering-failure-points-and-unhappy-paths-10f59fa6139a?source=rss-2fa4496fe40b------2</link>
            <guid isPermaLink="false">https://medium.com/p/10f59fa6139a</guid>
            <category><![CDATA[design-strategy]]></category>
            <category><![CDATA[systems-thinking]]></category>
            <category><![CDATA[service-design]]></category>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[resilience]]></category>
            <dc:creator><![CDATA[Greg Lakloufi]]></dc:creator>
            <pubDate>Thu, 11 Dec 2025 22:40:01 GMT</pubDate>
            <atom:updated>2025-12-17T14:13:54.503Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*K9a5twROnsC3YQljqo2FiA.jpeg" /></figure><h4>NASA and modern Service Design demonstrate that operational resilience is built by confronting <strong>mission-critical failure points</strong>.</h4><p>There was a time when people said the instant you have a plan B is the instant you’ve given up on plan A.</p><p>That line shaped entire generations of leaders. It rewarded focus, single-mindedness, and a belief in linear progress. It worked when the world had margins, buffers, and predictable cycles.</p><p>But that world is gone.</p><p>Today, clinging to a plan A without designing its corresponding unhappy path is not confidence. It is negligence.</p><p>In a global landscape defined by geopolitical tension, ecological disruption, digital fragmentation, and supply chain fragility, treating the happy path as the main path is not optimism. It is a structural risk disguised as strategy.</p><p>I’ve seen this firsthand throughout a decade of strategic and service design work. On every major project, the moments that defined the outcome were never on the happy path. They lived in the cracks: the failure states, the drop off points, the workarounds, the personal hacks, the moments when reality diverged from the immaculate journey map. The teams that succeeded were not the ones with the most beautiful plan A but the ones with the most resilient, well-designed plan B, C, D, and well, sometimes E.</p><p>This article is about why the unhappy path is now more important than ever, why ignoring it is dangerous, and how you must build it into the core of every strategy and decision you make.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*H3xpyK0bR_-5raDKLLHzlQ.jpeg" /></figure><h3>The Shiny Object Delusion: Why the Happy Path Is the Most Dangerous Path</h3><p>We are designing services inside a world that behaves nothing like the world most journeys were built for. Markets once followed patterns. Now they follow shocks. Geopolitical blocs realign, climate instability disrupts logistics, digital systems fragment across borders, supplier networks buckle, and cultural contexts shift faster than strategies can adapt.</p><p>Yet many leaders still design with the logic of a stable era. They optimize for best-case scenarios. They chase delight. They measure flow efficiency while ignoring failure thresholds. The result is a brittle ecosystem of experiences that collapse the moment something deviates from ideal conditions.</p><p>Every year brings a new technological promise, the next big thing destined to revolutionize the customer experience. And in the pursuit of the shiny object, companies overlook the far less glamorous realities that determine whether an experience survives contact with the real world. I have watched organizations spend months and years perfecting the ideal customer journey while leaving entire failure states unaddressed.</p><p>The patterns are painfully familiar:<br> • No recovery flow when authentication breaks.<br> • No redundancy when a logistics partner goes offline.<br> • No alternative journey when a digital channel fails.<br> • No operational slack to absorb a surge or shock.</p><p>This is not innovation. It is building on sand.</p><p>Whenever our teams lead through future state experience blueprinting, the breakthrough moment rarely comes from refining the happy path. It comes from surfacing the unloved steps, the off-ramps, the contingencies no one planned for. Those are the places where real-world strategy begins.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uB0_6AybpF-E8Kc9boDOKw.jpeg" /></figure><h3>How a Single Failure Point Can Collapse an Entire Experience</h3><p>A missing unhappy path is not an inconvenience. It is a systemic weakness. When failure states are not designed, a single breakdown can trigger a chain reaction. Trust erodes. Operations cascade. Staff improvise. Customers churn. Regulatory exposure increases.</p><p>A global telecom project I led became the clearest illustration of this. The company had engineered a beautifully choreographed experience made of 138 journey touchpoints. Months of design, research, operational planning, and orchestration went into weaving those physical, digital, and virtual steps into a seamless flow. But the entire experience depended on touchpoint number four: An automated email containing a validation code the customer needed to present in the retail shops.</p><p>When that email failed to arrive, everything came apart. Fast.</p><p>Technical errors, network delays, spam filters, or simple human error could all block that message. And when it did not arrive, the remaining 134 touchpoints instantly became utterly useless. Retail staff had no fallback protocol. Customers had no alternative path. The system had no redundancy. A multimillion-dollar customer experience hinged on a single fragile interaction. The unhappy path had never been addressed nor designed, and the experience collapsed under the weight of its own optimism.</p><p>Unfortunately, this is not an anomaly. It is how most systems are still built today.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*G15lTeX3nZxMp_JZNybnsg.jpeg" /></figure><h3>How NASA Designs for Worst Case Scenarios</h3><p>The Webb Space Telescope succeeded not because everything went right but because NASA designed for everything that could go wrong. Teams identified more than <a href="https://www.opb.org/article/2022/07/17/james-webb-telescope-had-344-single-point-failures-before-launch-then-success/">three hundred potential failure points</a>, every single one capable of ending the entire mission with a sobering loss of $10 billion invested over two decades of development. Every hinge, latch, mirror segment, and sunshield membrane had a defined contingency strategy. Webb launched to a location in space beyond the reach of any repair mission, making it essential to engineer solutions for every possible scenario in advance. The unhappy path became the design philosophy.</p><p>As one NASA engineer put it, “Prior to launch, we had 344 single-point failures. A single-point failure meant if this one thing fails, we could potentially lose the whole mission.”</p><p>This is the mindset of an organization that treats fragility as a fundamental reality. When the margin for error is measured in microns and mission lifetimes, every scenario must be examined and resolved long before the system leaves the ground.</p><p>Webb did not succeed because the happy path held. It succeeded because the unhappy path was engineered with absolute rigor. While many companies chase seamless delight, NASA confronts seam-full complexity. Their work proves that resilience is not luck. It is the result of engineering for failure before it happens.</p><p>NASA embraces what most organizations resist. In its engineering culture, the idea that failure is an option is not a reckless slogan but a mandate for preparedness. Every mission is a stack of risks, interdependencies, and unknowns. The only irresponsible approach is pretending otherwise. In NASA’s cleanrooms and control centers, teams rehearse breakdowns repeatedly because they know resilience is built by confronting what can fail, not by assuming success.</p><p>This is the mindset modern organizations now need most.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*KOhogj7Dxgop7Z8CPsiXrw.jpeg" /></figure><h3>The Unhappy Path as a Strategic Asset</h3><p>Here is where the narrative turns. The unhappy path is not a fallback strategy or a pessimistic exercise. It is not a defensive posture. It is a source of intelligence. It reveals assumptions, exposes interdependencies, and forces clarity long before reality does. It is the place where the hidden architecture of a service finally comes into view.</p><p>For years, I have watched teams transform entire programs after a single unhappy-path weakness came to light. One revealed fragility can sharpen intent, reset priorities, and surface work that would otherwise remain invisible until it breaks something expensive or irreversible. The unhappy path is where real leadership shows up, because it forces teams to address what they would rather ignore and to invest in the potential safeguards that never truly feel necessary until the moment they are.</p><p>And the lessons become clear when you look at what breaks and what can still be saved.</p><h4>1. The Ticketing Meltdown: When Scale Turns Against You</h4><p>A global event organizer built a ticketing system optimized for peak demand on paper but not in reality. On launch day, authentication slowed, customers refreshed their screens, and the system interpreted the activity as suspicious. Thousands were locked out instantly. With no fallback channel, everything dead-ended. Call centers collapsed. Social sentiment exploded. Regulators intervened. What should have been a victory became a global case study in how a single unplanned failure state can devastate an entire brand moment.</p><h4>2. The Retail Pivot: When a Failure Becomes a Better Design</h4><p>On another project, a retail brand’s new appointment system collapsed repeatedly. Instead of treating the failure as a crisis, the design team treated it as feedback. Customers improvised. Small groups formed. Staff stepped out from behind the counter to guide people manually. Conversations unfolded. The team realized the human choreography mattered more than the digital precision. They rebuilt the experience around hosted arrival rather than automated scheduling. A failure became a breakthrough.</p><h4>3. The Banking Outage That Exposed a Hidden Truth</h4><p>A major bank once experienced a three-hour outage in its mobile app. Customers could not check balances or transfer funds. Panic spread. But the real collapse was internal: no one had rehearsed a communication protocol for service failure. Teams debated wording. Legal slowed decisions. Executives delayed acknowledgment. The unhappy path had never been designed, and the reputational damage continued long after the system recovered. The technical outage lasted hours. The trust outage lasted years.</p><h4>4. The Airline Grounding That Should Never Have Happened</h4><p>A leading airline grounded thousands of flights when a single legacy system failed to relay updated crew scheduling data. The entire network froze. Travelers slept on airport floors. Competitors capitalized. Analysts estimated hundreds of millions in losses. The unhappy path was not hidden. It was documented. But it had never been rehearsed, stress-tested, or prioritized. One outdated system held a modern global operation hostage.</p><h4>5. The Hospital That Turned Crisis Into Strength</h4><p>A hospital’s patient-intake system went offline during a regional emergency. Staff quickly shifted to manual intake processes they had practiced but never had to use. Clipboards appeared. Runners moved information between departments. Temporary triage stations formed. The process looked messy but worked flawlessly because the unhappy path had been designed, drilled, and owned. After the crisis, the hospital integrated these learnings into a new hybrid model that improved care long after systems came back online.</p><h4>Why These Stories Matter</h4><p>Each example reveals the same truth: Systems do not collapse because the happy path failed. They collapse because the unhappy path did not exist at all.</p><p>The difference between disaster and resilience is rarely the technology or the strategy itself. It is the preparation for the worst, the imagination to anticipate failure, and the leadership required to design for it. When organizations learn to work inside the uncomfortable space of potential breakdown, they uncover the insights that make everything else stronger.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5poBnDWQYjLbvzZ_fAM68A.jpeg" /></figure><h3>A Practical Framework for Building Resilient Experiences</h3><p>To design for reality instead of fantasy, you need a different architecture. One built on discipline rather than hope, clarity rather than assumption. A resilient service begins with a systematic view of what can fail and how the system must respond.</p><p>The core practices are straightforward:<br>• Map systemic pressures, not just steps<br>• Identify failure nodes at every touchpoint<br>• Design meaningful redundancy<br>• Create operational slack<br>• Test like an adversary<br>• Build sensing mechanisms<br>• Codify the choreography of recovery</p><p>This is the blueprint of resilience. It is not glamorous. It is not fast. But it is the only way to design services and experiences worthy of the world they must operate in.</p><p>And when organizations commit to this work, something powerful happens. Trust grows. Stability increases. Loyalty deepens. The brands that thrive in the coming decade will not be the ones with the sleekest journeys or the most seductive interfaces. They will be the ones that offer reliability in an unreliable world, the ones customers return to because they can count on them when things go wrong, not just when things go right.</p><p>Trust is no longer earned by perfection. It is earned by preparedness and recovery. It is earned by designing for the full spectrum of human and systemic experience. When you design the unhappy path well, you do more than prevent failure. You create a strategic advantage that is difficult to copy because it becomes embedded deep within the operating logic of the organization.</p><p>This is the work that turns fragility into strength. This is the work that builds resilience. This is the work that endures.</p><h3>The Real Plan</h3><p>The old saying claimed that having a plan B meant giving up on plan A. Today, refusing to build a plan B is the real surrender. It is surrendering to delusion. It is surrendering to fragility. It is surrendering to a worldview built for a past that no longer exists.</p><p>The world has changed. Our work must change with it.</p><p>If your service can only perform when everything goes right, it is not ready for the era we now live in.</p><p>It is time to design responsibly. It is time to design resiliently. It is time to place the unhappy path at the center of strategy, design, and leadership.</p><p>This is not the backup plan.<br>This is the real plan.</p><p><em>(Disclaimer: The companies, projects, and examples mentioned in this article are based on publicly available information and industry best practices. No proprietary or confidential information has been disclosed, and all references are for illustrative purposes only. This article does not represent insider knowledge or violate any non-disclosure agreements. Some AI tools could have been used for basic editing, research and visuals, but the insights and perspectives are entirely human.)</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=10f59fa6139a" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>