How Do Design Leaders Differentiate in the Age of AI?

Design leaders have always been judged on the product: how it looked, how it performed. The team you built counted, and in stretches so did maturing the practice and earning design real influence. But under all of it, the product stayed the scorecard. And that rested on one thing. You could make the product, or lead the people who could. Now AI has ingested much of the making. The design system is training data. Anyone can prompt your patterns, and the artifact your title used to guarantee shows up in seconds, from anyone. Which leaves every design leader who came up before the shift with the same quiet question: when the machine makes the output, how do you differentiate?

You can read how the field is differentiating in their LinkedIn headlines, each one a wager on what makes that leader valuable. The wagers divide first into two kinds: you can bet on where you’ve been (ex-Google, ex-Stripe), or on what you’re for (the leader who raises the bar). Most bet on where they’ve been. The rest, betting on what they’re for, split into three camps.

Craft is the bet on judgment: the discernment, not the deliverable. The taste that knows which of ten good options is the right one, and can say why. A brief is a prompt now, and judgment travels only as far as you can put it into words.

Process is the bet on how the work runs. Leadership, strategy, scaling teams, zero to one, the operating system of a design org. The person who makes the machine go.

AI literacy is the newest bet. Fluency with the tools that rewrote the job: agentic workflows, prompting as a discipline, shipping AI-native product. The person who already lives in the new world.

I’ve been tuning my own positioning, so I wanted to see which bets my peers are actually placing. I pulled the headlines off 2,000+ of my LinkedIn connections (design leaders, product leaders, founders, the exact population living this) and classified each one for craft, process, and AI-literacy signals. A headline can carry more than one, so the shares overlap.

Process is the most crowded of the three. About 35% stake their headline on leadership, strategy, scaling, teams, operations, far more than claim craft or AI.

AI literacy is the newest signal, at around 16%. The tools it names barely existed two years ago, so almost no one was leading with it.

Craft is almost invisible: 2.5%. Fifty people out of nearly two thousand.

More than 55% lead with only a title and a company.

That 55% is really two groups wearing the same headline. When the company is a name, the logo is the flex. Most just never touched it. “VP of Design at [Company]” is what LinkedIn writes for you if you don’t. Either way, the bet isn’t on craft, process, or AI. It’s on pedigree: where you’ve been, which may or may not track what you can do. Five years ago “ex-[brand]” was the first filter a recruiter ran, and it was enough. Getting into those companies was hard, so the logo meant you’d already cleared a bar someone else set high. But a cleared bar is still a proxy, not the work itself, and AI just made the artifact cheap enough that the logo stops vouching for much.

That's the supply side. Demand tells a different story. I classified 42 design-leadership job descriptions, VP, Head, and Director, posted in the last month, for the same three signals. A job description lists everything it wants, so most postings name more than one signal, and the shares overlap. Process language shows up in all 42. Craft, quality, world-class, attention to detail: 30 of them, 71%. AI: 28, 67%, and climbing in the newest postings.

Set the two side by side and the gap is the story. Employers name craft as a requirement in seven of every ten roles. Design leaders make it their headline in one of forty. The thing the market keeps hiring for is the thing almost no one advertises. And leaders under-claim across the board: AI is asked for by two-thirds and led with by one in six, process by everyone and led with by a third. Craft is simply where the gap is widest.

Process isn’t just the most common of the three, it’s most of the field. And it’s crowded for an honest reason: process used to be the assignment. For most of the last fifteen years the market asked for process out loud, and craft only in the fine print. Up to about five years ago the ask was operating models, scale, the machine. For the seven to ten years before that it was design process, org stability, real integration with product and engineering. Build the function, mature it, earn the seat, make design legible to the business. So we did.

The people leading with “scaling teams” today are describing the job they were hired and promoted to do for a decade and a half.

Then the rules changed. The velocity playbook commoditized speed, and the work of running the org is next. Most of it is coordination, and coordination is exactly what these tools are getting good at. Process was never a cowardly bet. It’s just that the market spent fifteen years asking for it and then, only recently, quietly decided it was table stakes.

The 16% who lead with it aren’t wrong. But a differentiator everyone is racing toward might have a short shelf life. When every PM has Claude open in a tab and every designer prompts their way to a third iteration, “AI-native” reads the way “proficient in Figma” reads now: table stakes, not an edge. AI literacy is worth having. It just won’t be worth saying for much longer. Though I wonder how much longer.

Craft is the axis AI walks onto last. What it can’t bring is the discernment. Give a model the right prompts, the loops, your design system, and it produces the work. What it can’t tell you is how good the work needs to be. It doesn’t know the difference between good, great, and excellent. More to the point, it doesn’t know when each is the right call. That isn’t in the pixels, or the metadata. It’s a judgment about stakes and resources. Chase excellence where good would do and you’ve spent your attention on the wrong thing. Ship good where it had to be excellent and you’ve lost the users. Knowing which is which is the discernment. NN/g’s 2026 outlook lands in the same place: the durable value is judgment, not artifact production. It may be the most defensible thing a design leader owns.

And yet almost no one claims it. The most defensible position in the field is the one the fewest people name. It goes unclaimed because taste is the one thing a design leader can’t cleanly prove, especially in a headline.

I checked design leaders at the AI-forward companies everyone’s watching (Shopify, Deel, OpenAI, Ramp, Vercel) to see if they position differently. Two patterns, and craft is quiet in both.

The strong-brand crowd goes minimal. “VP of Design at Shopify.” “Head of Product Design, OpenAI.” A few just say “Designer.” When the logo is the proof, they let it carry everything, craft included. Implied, never stated. The builder crowd goes the other way and shows receipts. The sharpest headline in the whole search belongs to a design lead at Deel: “Shipped the first GPT-4 product · $3.2B GMV · 89% design system adoption · Scaling AI-native product teams.” That’s AI literacy and process, proven with numbers. Still no craft claim.

The exception proves it. One Shopify director writes “Craft, coherence & measurable outcomes.” That’s the whole move in four words: craft, named, then tied to an outcome so it doesn’t read as vanity. It stood out precisely because nearly no one else does it.

The two things a design leader gets hired to do, over and over, are “define what good looks like” and “raise the quality bar.” Those aren’t process asks. They’re craft asks: taste and judgment pointed at a whole org instead of a single screen. That’s the 71% of job descriptions talking: the market names craft in plain language, and the leaders who have it answer with process.

Ask why leaders answer a craft ask with process, and you hit an older habit. For fifteen years the climb to senior ran through the language of business. You proved you’d outgrown the pixels: you scaled the team, ran the operating model, earned the seat. Craft got filed as the junior move, the thing you did before you grew into leadership, and business fluency was the grown-up one. So when a job description asks for craft, the reflex is to answer in process, because process is the dialect that got us promoted. The catch is that the translation now buries the signal the market is scanning for. And it was never really a choice. Craft doesn’t compete with business fluency; expressed as an outcome, it becomes business fluency, the same judgment the ladder always wanted, finally legible in the language it speaks.

The taste was never the bottleneck. The execution was.

Those asks also used to be brutal to deliver. Defining good meant a design-system rebuild that ate multiple quarters. Raising the bar meant a brand overhaul, a season of roadmap tradeoffs, months spent getting forty people to agree on what “quality” even meant and then dragging the work up to it. The taste was never the bottleneck. The execution was. A design leader could see exactly what good looked like and still wait a year to watch it show up across the product.

That gap just collapsed. The same tools flattening everyone’s velocity also flattened the distance between a design leader’s judgment and its rollout. The system you spec gets rebuilt and propagated in weeks, not quarters. The bar you define gets encoded (examples, patterns, prompts) and pushed across every team at once, instead of living in a deck nobody opens.

If you lead design, your headline is already making one of these wagers, whether you chose it or not. Most of us are still running a default: the language the market used to reward, or a four-noun stack that hedges every bet at once. Neither is really a decision. Not anymore.

I’m not here to tell you craft is the answer. I’m still tuning my own positioning, and anyone selling a formula for a market this unsettled is guessing. What the data says is narrower and more useful: there are three lanes, and they aren’t equally open. Process is packed. AI literacy is filling fast. Craft, stated plainly and tied to a result, not left to the title to imply, is nearly empty. That’s not a prescription. It’s just where the room is.

The few who stake craft and make it land don’t reach for adjectives. They reach for something specific: an outcome, a sharp POV, a true detail no one else would write. Not “passionate about craft” but the taste call that killed three features and lifted activation. The process language, when it’s there, is evidence, not identity. And an outcome does one more thing: it clears the machine. The headline is the highest-weighted field in LinkedIn’s search, and that search now reads for meaning, not keyword matches. So an outcome carries the signal a recruiter’s AI is scanning for and the judgment a human is looking for.

Across the whole set, the headlines that stuck with a human, not a scanner, weren’t the most complete. They were the most specific. A design exec at PayPal lists “unlicensed AI moral psychologist • I fix $5 APIs,” and you remember it because no one else would write it. Another is just “designing with machines,” three words that stake a whole POV. A leadership coach leads with “I teach designers how power actually works,” a claim with a point of view instead of a category. None of them reads like a four-noun stack, which is exactly why they surface. Set them next to “Product Design Director | UX Leader | GenAI & Emerging Tech | Strategy, Systems, and Hands-On Craft” and the difference is physical: one is a person, the other is a job board.

The panic in this field has always been about what AI can do. The more interesting question, and the one I’m still sitting with, is what it can’t do.

The artifact was never the craft. It was the receipt. Now the receipt is free, and the judgment behind it has to stand on its own: how good the work has to be, and what’s worth building in the first place. The machine took the making. The deciding is still ours, and it was always the job. So claim it in your headline. The lane is wide open.

Sources: Nielsen Norman Group, State of UX 2026; UX Design Institute, UX Job Market 2026; my own [No Water Breaks](/ideas/no-water-breaks) and [If Design and Code Are Easy](/ideas/if-design-and-code-are-easy-whats-left-to-optimize). Network data: analysis of 1,963 LinkedIn connection headlines, classified across craft, process, and AI-literacy signals, July 2026. Company cohort: LinkedIn people-search of design-leadership titles at Shopify, Deel, OpenAI, Ramp, and Vercel, July 2026 (directional sample, not exhaustive). Demand side: 42 design-leadership job descriptions (VP / Head / Director of Design) posted in the 30 days to July 2026, coded for the same craft, process, and AI signals.