<?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 Ashish Chaudhary on Medium]]></title>
        <description><![CDATA[Stories by Ashish Chaudhary on Medium]]></description>
        <link>https://medium.com/@ashishdesignux?source=rss-22ebbb922a4a------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*aEW-LkxkViSrTMaemihQ_w.jpeg</url>
            <title>Stories by Ashish Chaudhary on Medium</title>
            <link>https://medium.com/@ashishdesignux?source=rss-22ebbb922a4a------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 23:49:07 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@ashishdesignux/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 “What If” Advantage]]></title>
            <link>https://medium.com/@ashishdesignux/the-what-if-advantage-4364778d7833?source=rss-22ebbb922a4a------2</link>
            <guid isPermaLink="false">https://medium.com/p/4364778d7833</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[future]]></category>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[ux-design]]></category>
            <category><![CDATA[design]]></category>
            <dc:creator><![CDATA[Ashish Chaudhary]]></dc:creator>
            <pubDate>Mon, 20 Jul 2026 19:05:00 GMT</pubDate>
            <atom:updated>2026-07-22T14:45:51.002Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*mCnJKNhailnoJfShn3nlzw.jpeg" /><figcaption>Created by author using Procreate and Gemini</figcaption></figure><h3>Why designers must un-standardize themselves in the age of AI</h3><p>Close your eyes for a moment and picture a website. Any website, let’s say a startup’s homepage.</p><p>Let me guess what you’re seeing. A navigation bar across the top. A big bold headline in the center. A subheading in grey. A rounded button, probably blue that says “Get Started.” Three columns with little icons. And somewhere, a smiling person looking at a laptop.</p><p>How did I do?</p><p>I didn’t read your mind. I read the industry’s mind. For the last fifteen years, we designers have been building the same website, the same app, over and over. Different logos, same design. And we didn’t do it by accident, we did it on purpose. We called it best practice. I want to convince you that this best practice has quietly become our biggest liability, with the advent of AI design tools.</p><p>I also want to show you the way out. Strangely, it doesn’t run through more technology. It runs through play, material, and speculation — the parts of design we were told weren’t “serious work.”</p><h3>How design got McDonaldized</h3><p>In 1993, the sociologist George Ritzer wrote <em>The McDonaldization of Society</em>. His argument: modern institutions organize themselves around four principles: efficiency, calculability, predictability, and control. A Big Mac in Tokyo tastes like a Big Mac in Toronto. That’s not a bug; it’s the business model.</p><p>Now read those four words again: efficiency, calculability, predictability, control and tell me that’s not a job description for a UX designer in the 2020s. We built design systems for predictability, ran A/B tests for calculability, adopted heuristics for efficiency, and standardized patterns for control.</p><p>And it worked. The web of the 1990s was chaos, blinking text, mystery navigation, Comic Sans overload and then human-centered practice came along and made web humane and usable for billions. I’m not here to diss usability.</p><p>But something else happened along the way, and it’s measurable. Researchers at Indiana University analyzed the visual evolution of thousands of websites across nearly a decade and found that web design has been steadily converging layouts, colors, and structures becoming quantifiably more similar year over year. The homogenization isn’t a vibe. It’s in the data. We got a web that is easy to use and impossible to tell apart. The Big Mac web.</p><p>I noticed this convergence from an unusual vantage point: as an immigrant to the discipline. I came to UX after a decade of practicing industrial design, and when I arrived at the University of Washington for my master’s in Human Centered Design &amp; Engineering, I was surprised and genuinely impressed by how academically evolved UX education had become. But it was a very different education from my Industrial Design masters at the National Institute of Design in India. NID’s pedagogy descended from the Bauhaus, and it put exploratory design making in order to discover on equal footing with problem solving. At UW, the center of gravity was HCD principles, applied impeccably. This is purely my observation, but I couldn’t shake the feeling that the system was gearing brilliant students to produce another highly functional, highly usable, highly predictable app. One school taught me to ask what a material could become; the other, what a user needs. We’re about to find out which of those questions a machine cannot answer.</p><p>Here’s the other irony that keeps me up at night. Do you know where UX got the very idea of “design patterns”? We borrowed it from architecture from Christopher Alexander’s 1977 book <em>A Pattern Language</em>. But Alexander never intended patterns as templates. He called it a <em>language</em> because he meant it generatively, the way words combine into sentences that have never been spoken before. We took the vocabulary and threw away the poetry. We turned a language into a stencil.</p><h3>Why this is now an emergency</h3><p>For twenty years, “designers who follow patterns flawlessly” had a great career and success in design was defined by the ability to master specific patterns. As intensive programs began teaching these skills in twelve-week cycles, it signaled an industry shift: the tasks that rely on repetitive pattern-matching are becoming increasingly accessible to automated tools, allowing designers to focus on higher-level problem solving.</p><p>Think about what a large language model fundamentally is: a system trained on everything that has already been made, extraordinarily good at producing the most statistically probable next thing. And what did we spend two decades doing? Making our discipline as statistically probable as possible converging, documenting exhaustively, publishing millions of nearly identical examples for the machines to study.</p><p>We didn’t just build the McDonald’s of design. We wrote the training manual and left it on the counter.</p><p>So the question you should be asking isn’t “will AI take my job?” It’s: <em>which part of my job was only ever pattern-retrieval and what’s left of me when that part is gone?</em> What’s left, I want to suggest, is not a smaller thing. It’s the original design before we industrialized it.</p><h3>“What if” before “How might we”</h3><p>You know the sacred phrase of Design Thinking: “How might we…?” It’s a wonderful question. But notice that it assumes that the problem is already known, already framed, already handed to you in a ticket. “How might we” is an optimization question, and optimization is precisely the game AI plays better than us.</p><p>There’s an older, wilder question that comes before it: <em>What if?</em></p><p>What if the interface disappeared? What if your data had a smell? What if a meeting could feel tired? “What if” doesn’t optimize a known problem, it invents the problem space. And inventing problem spaces is, so far, a stubbornly human sport, because it requires wanting a world that doesn’t exist yet.</p><p>Anthony Dunne and Fiona Raby called it <em>speculative design</em> design that asks questions rather than only solving problems, opening up possible futures for debate before we’re trapped in one of them. James Auger called the craft a “perceptual bridge” between the world as it is and as it could be: too close and it’s boring, too far and it’s science fiction.</p><p>And where did the genuinely new paradigms come from? Not from usability-testing the old ones. Multi-touch, the graphical interface, the Dynabook that became your tablet all began as speculations, demos, toys. As Alan Kay put it: “The best way to predict the future is to invent it.” What-if sits upstream of how-might-we, and upstream is where the leverage lives.</p><h3>The non-UCD toolkit</h3><p>How do you practice “what if”? By recovering some traditions used in industrial design and architecture that are less familiar to UX designers, traditions where the user is <em>not</em> the starting point, and that’s exactly the point.</p><p><strong>Design through play.</strong> Imagine you have a piece of paper in your hand, you’re just playing with it without a goal. Bending it, folding it , crumpling it, having fun with it, and it may produce a form which suddenly looks like a cool idea for furniture or lamp or just something decorative if translated into a different material. This is exactly what Paul Cocksedge did and created this fascinating piece called Poised Table.</p><figure><img alt="A folded sheet metal table called Poised Table" src="https://cdn-images-1.medium.com/max/1000/1*3GBUTErzT-00A1haNAFypQ.jpeg" /><figcaption>Poised table by designer Paul Cocksedge</figcaption></figure><p>This is the core of ‘Design through play’, where your starting point is not a problem that needs a solution rather an activity that induces imagination, and creative experimentation which later can be reflected back upon and (maybe) converted into “something”.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*igiTQjyR_tABZ5f1yPy0dw.jpeg" /><figcaption>Photo by <a href="https://unsplash.com/@ssdriscoll?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Steve driscoll</a> on <a href="https://unsplash.com/photos/gray-architectural-building-VsBl5PwVZpY?utm_source=unsplash&amp;utm_medium=referral&amp;utm_content=creditCopyText">Unsplash</a></figcaption></figure><p>Using material as a starting point of exploration and wonder is a key part of ‘Design through play’, and sometimes also called ‘Design through material’.</p><p>I spent ten years as an industrial designer before I ever designed a screen, so I know in my heart what Donald Schön meant when he described design as “a reflective conversation with the materials of a situation.” The material talks back; skill is learning to listen. Frank Gehry didn’t ask “how might we optimize museum wayfinding?” He played with crumpled paper and titanium, then dragged aerospace software into architecture because no architectural tool could describe what his hands had found. The result was the Guggenheim Bilbao, a building so transformative that economists named an effect after it.</p><p>A ton of interaction design paradigms have emerged not because someone was chasing a problem to solve but rather was trying the limits for given technology. The best example of this is a feature launched in Google I/O “Select from Screen”. The HCI researcher who initiated this idea told me, he was curious about what if, AI meet the way people interact with each other. This curiosity led him learning about people and how AI can bend to meet the current way they interact with each other. He worked for months to test the limits of the AI modal he was working with, and it felt like he was treating AI not as a tool but as material that he could fold, bend and crumple to discover new interactions.</p><p>And here’s the thing I most want you to hear: <strong>Treat AI as a material to play with.</strong> It’s the strangest, most expressive design material we’ve ever been handed, probabilistic, conversational, occasionally hallucinating. You can’t understand it from guidelines, because the guidelines are being written right now by the people willing to get their hands dirty.</p><p><strong>Speculation as rehearsal.</strong> AI is trained entirely on the past; it interpolates within what has been. Speculation is the deliberate practice of stepping outside the training distribution. It is, quite literally, the one move the machine cannot make on its own, you cannot retrieve a future that has never been written down.</p><h3>What I’m asking you to do</h3><p>While scholar from the Netherlands, who penned <em>Homo Ludens: A Study of the Play-Element in Culture</em>, Jojan Huizinga, characterized play as a foundational element of civilization, preceding even culture itself. It represents an inherent human drive, a primordial instinct that requires no formal education. However it’s easier said than done in a “typical work” environment.</p><p>What if is hard to practice, especially in a traditional corporate structure, where design teams act mostly as guardians of the design patterns. The mandate of product orgs reward step changes over radical new ideas because of the “product maintenance” mindset, and most of the new ideas or slightly radical ideas land in P3- P4 in MVP documents and die there silently.</p><p>My opinionated stand is, this might change very soon and I will elaborate on this in my next article.</p><p>This article is not a call to abandon User-Centered Design. Usability is table stakes; accessibility is a moral floor. Ship the predictable checkout flow better yet, let AI ship it.</p><p>I’m asking you to re-sequence yourself. Protect a “what if” before any project gets its problem statement. Tinker with the material weekly, build for one, before shipping for millions. And build one useless thing per quarter, because Eames was right: toys are the only reliable preludes to serious ideas.</p><p>At the start of this essay, you closed your eyes and saw the same website everyone else saw. That was our report card. Patterns are memories, and you cannot out-remember a machine that has memorized everything.</p><p>You can, however, out-imagine it.</p><p>So the next time you sit down to design, maybe don’t start with the user, the guideline, or the ticket. Start with the two most dangerously human words in our language:</p><p>What if.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4364778d7833" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The golden era of idea economy and the rise of dreamers!]]></title>
            <link>https://medium.com/@ashishdesignux/the-golden-era-of-idea-economy-and-the-rise-of-dreamers-7720f41c6d58?source=rss-22ebbb922a4a------2</link>
            <guid isPermaLink="false">https://medium.com/p/7720f41c6d58</guid>
            <category><![CDATA[positive-thinking]]></category>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[makers]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Ashish Chaudhary]]></dc:creator>
            <pubDate>Tue, 30 Jun 2026 18:58:24 GMT</pubDate>
            <atom:updated>2026-06-30T19:14:13.375Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="an abstarct illustration of human figures reaching out and finding their star (AI generated)" src="https://cdn-images-1.medium.com/max/1024/1*1MzEmTPTQCwu5cEctqtHCw.jpeg" /><figcaption>Made with AI</figcaption></figure><p><em>Disclaimer: These are my personal opinions and do not reflect my employer’s views.</em></p><p>A Sunday prototype gathers ten thousand active users before securing a co-founder. A functional demo commands capital before generating a presentation slide. Solo engineers ship software, watch unfamiliar cohorts grow dependent on it, and only then determine whether a business exists. This is more than a modern trend; it is the total collapse of the traditional <em>genesis-to-scale</em> cycle. Observe the current velocity of bespoke AI wrappers and autonomous tools, the swift ascension of vibe coded projects or lone creators populating the indie-SaaS landscape. Figures like Pieter Levels have reduced this to a pure methodology: fabricating single-use applications in hours, broadcasting them via minimalist media previews on networks like X, and capturing thousands of monetized accounts before drafting a visual brand. From the rapid expansion of ElevenLabs to obscure operators launching automated architectures from private spaces, the current ecosystem proves that genuine validation occurs long before legal registration is ever completed.</p><p>And these are the signs of a brave new era, The age of dreamers: powered by agentic technology!</p><h3>What the idea economy actually is</h3><p>We are living in a creator economy. The rise of social media platforms like TikTok, Instagram are testament to the idea of content creation as a new “job”. The creator economy monetized attention. You made something watchable, an audience formed, and the audience was the asset.</p><p>Now with the rise of AI we are moving towards the idea economy which monetizes possibility. You imagine something useful, strange, or delightful enough that it pulls demand toward itself, and the demand is the asset.</p><p>The difference is what the work produces. A creator makes content for an audience that already exists. A dreamer produces <em>executable concepts</em>: prototypes, demos, half-built products. Things that didn’t exist yesterday and may not have a market until they do.</p><h3>Why now</h3><p>Execution used to be the tax on imagination. To build, you needed engineers. To ship, you needed capital. Execution acted as a filter, and it was never a neutral one. It favored people with access, credentials, money, and geography. Millions of ideas died because the people who had them couldn’t afford to try.</p><p>AI weakens that filter. The first 80% of building software is becoming a utility. The last 20% (reliability, security, distribution, trust) is not. Hold that thought.</p><p>People who could imagine but not build can now build. People who could build but not imagine just lost their moat. The bottleneck is moving to favor dreamers.</p><h3>AI gives imagination a mirror</h3><p>The most important thing AI gives dreamers isn’t perfection. It’s first contact with their own ideas. <em>Show me the interface. Write the pitch. Make it stranger.</em> Many ideas only become clear after they become visible. A dreamer can build something rough first and use that artifact to create belief.</p><p>But vague ideas stay worthless. <em>An AI travel app.</em> <em>A better calendar.</em> Not enough. The valuable dreamer has a high-resolution<strong> imagination</strong>.</p><p>They don’t just say: <em>I want to build a travel app.</em> They say: <em>I want to build an AI travel companion for families planning trips with young children, where the system understands budget anxiety, vegetarian food constraints, jet lag, elder accessibility, and the emotional desire to create memories before children grow up.</em></p><p>That’s not an idea. That’s a world. AI rewards that specificity. Average prompts produce average outputs. The future belongs to the most disciplined imaginations, not the most numerous.</p><h3>The pathologies nobody wants to name</h3><p>Most essays about this shift stop at the celebration. That’s where the real story begins.</p><p>When execution becomes cheap, bad execution becomes cheap too. More spam. More derivative products. More things that look polished and feel hollow. Output is not meaning.</p><p>The idea economy has its own failure modes:</p><p><strong>Idea inequality.</strong> Dreamers with better tools, networks, and access dominate. The old gap reappears in new clothes, wearing the language of democratization.</p><p><strong>Homogenization.</strong> When everyone prompts the same models trained on the same corpus, outputs converge. The most valuable dreamers will be the ones who bring the weirdly specific, the culturally particular, the emotionally strange.</p><p><strong>Speculative bubbles.</strong> When ideas become investable before they’re proven, capital chases the flashy and the hollow. Blitzscaling’s hangover (MoviePass, WeWork) is what happens when ‘<em>can we build it</em>’ passes and <em>should it ‘exist</em> fails’.</p><p><strong>IP protection.</strong>Current intellectual property frameworks are fundamentally unequipped for this shift. Historically, IP law protected the <em>expression</em> of an idea (like a specific painting or a written script), not the idea itself. But when AI can bridge the gap between a fleeting thought and a fully realized digital asset in seconds, that legal distinction completely collapses.</p><p>These aren’t future risks. They’re the shape of the present.</p><h3>What deserves to exist</h3><p>When you can build anything, the question shifts from <em>can we build it</em> to <em>what deserves to exist</em>. That demands sharper lenses than feasibility:</p><ul><li><strong>Desirability.</strong> The best ideas surface ‘wants’ people haven’t named yet. Airbnb didn’t solve a housing problem. It revealed that millions of travelers wanted to live like locals, a desire most couldn’t articulate until the product existed.</li><li><strong>Timing.</strong> General Magic shipped a touchscreen personal communicator in 1994. There was no mobile internet, no app store, no ready audience. The company folded. Its alumni built the iPod, iPhone, Android, eBay. When building is fast, reading the moment matters more, not less.</li><li><strong>Sustainability.</strong> Does the idea create value or extract it? The answer matters before the round closes, not after.</li><li><strong>Delight.</strong> When everyone ships functional, the winners make people feel something. Apple invented none of the categories it owns. It made each object desirable.</li></ul><p>Use these as a rubric, not a vibe.</p><h3>The moats that are left</h3><p>In my last <a href="https://medium.com/@ashishdesignux/the-last-designer-in-the-room-28f5515de9a0">article</a> I argued that taste is <em>the</em> moat now. Taste is <em>a</em> moat but not the last one, there are others, and pretending otherwise will get dreamers killed by incumbents.</p><p><strong>Distribution.</strong> Large organizations with a huge distribution advantage over which allows them to reach out to a large user base,quickly. The post-execution moat is earned access to an audience that already trusts you.</p><p><strong>Trust and data.</strong> The last 20% of real software (reliability, safety, the proprietary signal from operating at scale) is where durable advantages live. Gatekeeping isn’t bad in all cases. It allows the systems that don’t fall over and users don’t get hurt. Any honest account has to grapple with this.</p><p><strong>Taste and judgment.</strong> Necessary, insufficient on their own. The winners pair taste with the unglamorous work of earning distribution and trust.</p><h3>The upward movement of ‘expertise’.</h3><p>Designers spend less time producing every artifact and more time defining quality and curating outputs. Developers spend less time writing every line and more time on architecture and reliability. Dreamers move downward into execution. Experts move upward into judgment. The middle gets compressed.</p><p>A dreamer can now bring a vivid prototype to a designer or engineer much earlier. The expert doesn’t start from a blank page. They critique, refine, strengthen, scale. The best experts won’t be gatekeepers of execution. They’ll be amplifiers of high-quality imagination.</p><h3>The dreamer’s discipline</h3><p>This is a homecoming for a specific tribe: the people with an “Ideas for Later” folder, the keepers of barely-legible diagrams. For years they were told to get back to work. In an age of agentic execution, those scribbles aren’t distractions. They’re blueprints.</p><p>But AI won’t reward all dreamers. Three moves matter most:</p><ul><li><strong>Be weirdly specific.</strong> AI is trained on patterns. You win by bringing what it doesn’t have: obsession, lived context, emotional texture. <em>Make it feel like a children’s science museum from 1989</em> beats <em>make it fun</em>, every time.</li><li><strong>Build before convincing.</strong> The old world required a pitch before a prototype. Now you can have a demo before anyone asks for one. Use that leverage.</li><li><strong>Learn to kill your own ideas.</strong> Starting is easy. The hard part is knowing what deserves more life, and what should die.</li></ul><p>Without taste, dreamers make noise. Without judgment, they mistake novelty for value. Without ethics, they get dangerous.</p><h3>This is the time…</h3><p>Take out the old notebooks, mails you sent to yourself in the middle of the night with a subject “idea” or a feature that was earlier deemed P4 by the team. <br>Make them happen.</p><p>The wall between imagination and reality is shorter than it has ever been.</p><p>But bring your sharpest judgment. Taste is necessary. It is not sufficient. The dreamers who win will pair imagination with the unglamorous work of earning distribution, trust, and the right to be heard.</p><p>The dreamer’s real superpower was never imagination.</p><p>It was knowing what deserves to exist, and being willing to do the hard part after.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7720f41c6d58" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Last Designer in the Room]]></title>
            <link>https://medium.com/@ashishdesignux/the-last-designer-in-the-room-28f5515de9a0?source=rss-22ebbb922a4a------2</link>
            <guid isPermaLink="false">https://medium.com/p/28f5515de9a0</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[design-in-ai-era]]></category>
            <category><![CDATA[ux-design]]></category>
            <category><![CDATA[changing-careers]]></category>
            <category><![CDATA[paradigm-shift]]></category>
            <dc:creator><![CDATA[Ashish Chaudhary]]></dc:creator>
            <pubDate>Wed, 18 Mar 2026 13:27:20 GMT</pubDate>
            <atom:updated>2026-03-18T13:27:20.616Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*E0mP_gebILb9NaKQ73ETpw.png" /><figcaption>Finding star by Ashish (Procreate + Seedream)</figcaption></figure><p><em>Disclaimer: These are my personal opinions and do not reflect my employer&#39;s views.</em></p><p>I’ve spent the last few years designing near-future AI experiences, the kind where you’re inventing interaction patterns for capabilities that didn’t exist six months ago. And I’ll tell you what keeps me up at night: it’s not whether AI will reshape the design industry. It’s whether designers will realize what’s actually changing before it’s too late.</p><p>The cost of knowledge work is declining. An AI agent can run competitive analysis in seconds, synthesize a thousand interviews overnight, and ship production code before lunch. When execution becomes a commodity, the boundaries between PM, UX, and Engineering don’t just blur; they dissolve.</p><p>So, what is left for the design industry? If the machine can make anything, the designer’s value shifts from <strong>what can be made</strong> to <strong>what is chosen to be made and why.</strong></p><h3>1. When you can build anything, the only question is: should you?</h3><p>Here’s the paradox: as the cost of execution drops, the value of execution drops with it. In the “old” world, friction was the designer’s natural filter. A practitioner couldn’t prototype fifty ideas, so they chose three. That act of choosing <em>was</em> the work.</p><p>When an AI agent can generate fifty product concepts by morning standup, the bottleneck moves upstream. The industry is moving from “Can this be built?” to the much harder: “Is this worth building?”</p><p>Designers have a unique edge here, but it isn’t what many think. While a PM evaluates through market viability and an engineer through technical elegance, a designer’s evaluation is <strong>experiential</strong>. Designers “feel” an idea forward. They simulate the moment of use, inhabiting the user’s confusion or delight before a single pixel exists.</p><p>AI is superb at extrapolation and extending existing patterns. But transformative ideas are discontinuities. The iPhone didn’t come from flip-phone data; Airbnb didn’t come from hotel surveys. Claiming that gap between “what is” and “what could be” is the new primary territory for the design profession.</p><h3>2. User advocacy isn’t about data anymore. It’s about values.</h3><p>The design industry needs to be honest: AI agents may soon have an unfathomable amount of user data. They can process every support ticket and session recording in minutes, work that used to take research teams months.</p><p>But user advocacy was never really about data collection. It was about <strong>moral priority</strong>.</p><p>When designers advocate for users, they are making an ethical choice: <em>“I will fight for user interests even when they conflict with the bottom line.”</em> An AI given a revenue-centric goal might see that a dark pattern lifts checkout conversion by 12% and keep it. It takes a human to decide to remove it because it erodes trust.</p><p>There’s another layer here. AI mostly processes user data literally — it reads what people <em>say</em>. Designers are trained to notice what people <em>don’t </em>say. The hesitation, the workaround that reveals a hidden need, the gap between what a user asks for and what they actually need.</p><p>In the agentic future, a designer’s role in user advocacy shifts from gathering insights to interpreting them. AI delivers the signal; the designer provides the meaning. That’s the difference between knowing users drop off at step three and understanding they drop off because the flow <em>asks for trust before it’s earned</em>.</p><h3>3. Taste is the last moat</h3><p>If AI generates the designs, what happens to craft?</p><p>AI-generated interfaces are already “good enough.” They follow patterns, maintain consistency, and hit accessibility standards. But “good enough” describes the floor, not the ceiling.</p><p>Think of two products with identical features. One is generated by an AI following best practices. The other is art-directed by a human with an opinion about spacing, typography, and that ineffable quality designers call “feel.” Users will pick the second one every time.</p><p>What they’re responding to is <strong>Taste</strong>. Taste is a departure from the average. It’s the decision to break a grid for emphasis or to slow an animation by 80 milliseconds to make it feel more intentional. AI can mimic patterns, but it cannot possess an aesthetic worldview.</p><p>Designers are being promoted from makers to <strong>Art Directors</strong>. Like a film director who doesn’t operate the camera but whose vision makes the film cohesive, the professional output is no longer Figma files. It is <em>direction</em>!</p><h3>How to thrive in the Agentic Era</h3><p>My advice for designers is simple: stop defending the tools. Figma, prototyping, and usability testing are all being automated. If a designer is only called in to “make it look good,” that battle is already lost. Here is how to pivot:</p><ul><li><strong>Master taste, not tools:</strong> Study architecture, film, and industrial design. Build a point of view on quality that exists beyond a software suite.</li><li><strong>Be a polymath:</strong> It’s about connecting the dots to get to novel solutions that don’t exist in the training data. AI synthesizes what has already been done; the designer’s job is to pull from psychology, history, and the physical arts to identify the gaps between existing data and future needs.</li><li><strong>Interpret, don’t just gather:</strong> Use AI to synthesize the “what,” but spend energy explaining the “why.”</li><li><strong>Get upstream:</strong> Move into the rooms where the “Should this be built?” questions are asked. The earlier a designer is in the process, the more leverage their judgment has.</li><li><strong>Practice saying “No”:</strong> When AI makes options infinite, the most valuable skill is knowing which ideas to kill.</li></ul><h3>The New Designer’s Playbook</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VMgYX46iLkUP_itQStVq_Q.png" /><figcaption>Old vs new role</figcaption></figure><p>There’s a seductive narrative that AI will do everything so that humans will do nothing. The truth is more interesting: AI may be able to do everything that can be <em>specified</em>.</p><p>But imagining what <em>should</em> exist, who it serves, and what quality it aspires to — that is precisely the work that resists specification. It lives in the designer’s values, taste, and the irreducible capacity to care about another person’s experience.</p><p>That is the territory the designers in the industry need to claim.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=28f5515de9a0" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>