“AI is taking us from throwing spaghetti at the wall to see what sticks, to getting a spaghetti throwing machine that can do it much faster.”
– Bob Baxley

A designer recently joined us and spent three weeks making all the right moves to onboard: she read the system, sat in on reviews, asked good questions. Then she asked what a flight result actually is, and someone opened Figma and pointed at a component in the library.

This is certainly an answer to what it looks like, but it really doesn’t say which states the thing can be in, what’s legal to do to it in each one, which contexts it belongs in, or what happens to it when the trip it sits inside gets canned.

We have an answer for each and every one of those somewhere in the building, but none of them written down, because since the dawn of UI design, the only thing that ever needed to read our design decisions was another designer, and they could just ask us, or use common human sense and experience.

And that’s about to become the most expensive gap in your organization.

Directive screens, and what’s going to replace them.

Everything we design today is directive, meaning to get something done you go to a page, fill in fields, press a button, then go to the next page. We have a design system to supply the parts those pages are built from, which is why component libraries are the artifacts we’ve spent a years perfecting.

Intent-aware interfaces flip that completely on its head. Instead, someone states a goal or asks a question, and the system then works out what they’re trying to do from the words they used, the account they’re using, what they’ve done before, and how precisely they phrased it. Then, through the magic of AI✨, it composes a surface for that intent instead of just routing them to a page someone drew in advance.

The real difference shows up in the shape of what gets composed. “Move all my October flights to November” is a bulk operation, so it should give something built for reviewing forty changes at once, with the new dates called out in a column or something. “Which route gets me there cheapest in November” is exploratory and fluffy, so it should give me something closer to a canvas or map, built for comparing and wandering. The same underlying data, the same design system, two completely different things materializing on my screen, neither of which exists as a page in my Figma file today.

We’re already doing this in some paces. Ask an AI assistant for a recipe and you get a card with a timer that starts when you tell it to, and all the ingredients neatly stacked by order of when you need them. Ask it to prioritize your week and you get a sortable list. Ask it to draft a message and you get an editable box with alternative ways to write something. No-one designed that recipe page, the priorities page or the drafting one. The system simply recognized three different intents and assembled three different surfaces for those.

The middle ground was always the ceiling

When we work with fixed screens, we’re forcing a compromise that we’ve gotten so used to, we stopped noticing it.

What people need varies enormously and going with a single layout can only serve a single point on that axis. On the first use a person wants guidance and focus. On the one hundredth, that same person now wants density, shortcuts, and maybe even settings. A traveler booking a single flight and an agent moving forty passengers around are using the same screen with wildly different needs. And since we haven’t been able to magically adapt at scale, everyone gets the average of all of them, which runs too slow for the expert and demands way too much precision from the beginner. You end up with something that takes the confusion of the power user interface, and the limitations of the beginner’s.

And we happily shipped that for decades and called it best practice, because there wasn't really an alternative. But that interface can now be assembled per person (and per moment), so the compromise has stopped being a constraint and turned into a choice instead. What stands between you and the alternative is that the thing doing the assembling has no idea what any of the things in your product are.

The layer no-one has written.

In Death of a WIMP Designer, I wrote about an example of this, and its solution. A model with complete access to Shopify’s Polaris library - every component and composition pattern they had ever documented - was asked to build two very different things: a confirmation for one small change, and a bulk edit across thirty records. It produced nearly the same layout twice. The library told it what a discount looks like, but never told it what a discount is for.

But when we add a second layer of documentation, the same model suddenly gives us a compact modal for one and a batch editing screen for the other.

That new layer is the whole ballgame, and almost no-one has it.

This semantic layer has three core parts to it:

First, your primitives, which are the abstract “things” people actually create and manipulate - these are your whats. It could be a flight result, a trip, a traveler, a fare alert, or a booking. Every product you’ll build has fewer than ten of them and yours have probably never been written down anywhere. Each one has its own little aura or signifier, meaning how a person recognizes it wherever it turns up, along with its essential fields, its states, the transitions and actions that are legal between them, and what it sits above and below it in the greater hierarchy.

Second, you’ve got your surfaces, which are containers rather than pages, or components. These are your hows - how the primitives appear and react to actions or data saturation, and make sense for the end user. A batch surface for many objects at once. A discovery surface for someone who doesn’t yet know what they want. Each one documented with its purpose and which primitives it’s meant to have squeezed inside it. These are essentially all the variants you can think of for what we call components today, but meant for a different consumer.

Third, you have intent signals, which is the part everyone seems to skip. A brand-new account opening trips for the first time signals learning and onboarding opportunities. A request naming a date range and forty records signals bulk work and “get out of my way”. The number of of “things” someone already has done or created signals their experience with the product. Write those down and the composition stops being a guessing game.

Objects then become the material.

With directive design, we place inputs at the center, because the fields and buttons were what people touch. When people stop touching them, the object itself becomes the thing you’re designing, and it now has to survive being rendered anywhere it pleases.

In my case, I’ll go and find every place a flight result currently appears. Search, comparison, trip planning, confirmation, price alerts, a fare-drop email, whatever our AI assistant renders when someone asks in plain language, and then put them in a nice long row row, which almost nobody does, since they were built by different people at different times against different deadlines and KPIs, and who’s got the time for this anyway? Some lead with price and some with duration, some have a button, some don’t. The layover is a number here and a graphic over there. The carrier logo appears on five of them and not the other three for some reason.

And don’t get me wrong - it’s not that the designers that did all these made a wrong call. The “thing” is incoherent across the whole product anyway, so it was never a quality problem that more review was ever going to fix. And when we juggle around with an object with no real true definition, it ends up being defined locally by every surface that needs it, and local definitions tend to drift. It’s like a game of telephone, but instead of a funny sentence at the end, we get crippling tech debt. When a human is drawing up all your screens manually, we can mitigate that, but when a machine is composing surfaces you never saw, this becomes the thing that makes or breaks.

Which means the objects (things) in our product need the same level of attention we currently reserve for our buttons and sheets.

So who owns which bit?

The split that makes this work in an org is like grammar against vocabulary, and we can map it onto teams we already have.

Design systems own the grammar. Which surfaces exist in this product at all, the metadata every primitive has to carry around with it, how objects nest inside each other, and the signifier (aura, fingerprint, whatever) system. Vertical (or problem space) teams then own the vocabulary. These are the fields, the business rules, the states, the relationships and contracts between the “things”.

When we combine them, the system can finally start composing sentences.

Chances are, in most organizations, this work belongs to nobody right now, because it’s too product-specific for the systems team and too systemic for any single product team, and it sounds really long-winded and heavy, and those A/B tests aren’t going to magically produce themselves. But someone’s got to grab it.

The cheapest place to start needs no AI at all. You take your most important object and check how it’s identifiable across every surface it jumps into, and then make it consistent. This is the crucial, but boring work the vibe bros don’t tell you about in their tutorials. You need to get your hows, whys, and whats into great shape before you can build understanding and agency in your system.

But, here’s the but.

Muscle memory is real, and any surface (UI) that quietly rearranges itself - however clever the it might be around that shuffle - is essentially hostile. The thing to zero in on is whether that adaptation answers something the person asked for, or if the system just kind of watched them and decided on its own.

And, truth be told, this is the part we’ve yet to figure out.

We can see the evolution happening, clear as day. That isn’t in question. Screens gave way to pages, pages are giving way to surfaces composed on demand, and the tools to do it are already in your stack. What we can and should question is whether the service composing them knows what your product is made of, or has to guess from the pictures you left scattered all over the place.

And the only one who can piece this together is you.

This article is the fourth in my AI series and a direct follow-up to part 3:

Hi, I’m Oscar - Founding designer at momondo, I’ve won a Material Design Award for Innovation, and I help design leaders succeed.

Subscribe now

Measurement is not neutral about what it rewards. It prefers whatever sits closest to hand, and in product the things closest to hand are the tangible ones like conversion, retention, time on task, revenue per user. Countable this sprint, so we manage it this sprint.

And the intangibles wait outside in the cold.

Whether the product met the need someone arrived with, or whether it felt trustworthy. Whether the promise it makes on the first screen is still true by the last. None of these resolve to a clean number, and each asks something of you before you can read it at all, like a clear picture of the product and an honest understanding of the person using it. A lot harder than opening a dashboard. So it waits, and waiting turns into never, and the countable half wins the budget argument every quarter until nobody is left maintaining the other half.

And then eventually a substitution happens that no one consciously decides, and the business metric stops standing in for the experience and becomes the experience. Performance becomes the whole universe, and what the numbers show is treated as the only reality. What they cannot show is treated as absent. A team can hit every target on the board and still ship something nobody wanted, because the board was never able to see the want in the first place.

It’s not new, and it’s not unsolved

Service designers spotted the problem decades ago, and SERVQUAL, built in the late 1980s, reads quality as the gap between what a customer expected and what they perceived, across five dimensions. SERVPERF followed a few years later, arguing you could drop the expectation side and score perceived performance on its own. Both are rigorous, and both have held up for forty years.

They are also pretty hard to live with as they were designed for service encounters and periodic surveys. A long questionnaire after the fact, dimensions to reconcile by hand, a research project every time you want an answer. Few product teams will be able to run that every sprint, so they run nothing, and soon enough they drift back to the dashboard already open in the next tab. The old models were right about the problem, but they were built for a slower world than the one a modern product team ships in.

What a modern product team needs is something lighter, quick enough to live inside a normal cycle, plain enough that brand, product and research can read it without a manual, and still honest about the intangible.

Three truths, three gaps, and a pizza place

To get a clear picture, we just have to look at three simple truths side by side: Need → narrative → reality.

First up, the need, which is what matters most to the person, in their terms, never the action you wish they would take. Then, the narrative is what you lead them to expect, and it runs far wider than marketing: a screen that lays everything out and lets you compare is already promising you are in control here, without a word written. Finally, the reality is then what the person concludes is true after using the thing. Not whether the feature worked, but whether they came away confident, or lost, or quietly sold to.

The gaps between them are where you find the clues as to what’s really going on. Between need and narrative you find relevance – whether the story speaks to something the person values. Between narrative and reality sits credibility – whether the experience backs up what you claimed. And between need and reality you have fulfillment – does the product resolve what the person came for, regardless of what you said about it.

Close all three gaps and something materializes in the middle that none of them produces alone: brand integrity. It is the compound of a promise made well and kept, and like anything compound it builds on itself, a little more each time a person wants something, believes you will provide it, and finds that you did.

Treat the intangible like the tangible

The model only works, though, when you actively dispel the asymmetry that started all this. When the need, the narrative and the felt reality get measured with the same care and the same weight you give conversion and revenue.

They earned that weight because they decide whether the hard numbers still hold next year.

Reality can be read, but it has to be read per person and per context, because the same delivery lands differently depending on who meets it. Show every option and a confident user reads control, a newcomer reads overload, someone in a hurry reads I can’t tell what matters. One average score flattens those three into a number true of nobody. But when you score them apart, against what each person needed, then the intangible stops being a feeling and turns into evidence you can act on.

Eight states, and where to point next

Each truth is either strong or weak, and that gives you eight combinations, and every single one points somewhere different.

Three strong ones show coherent value, the state you want to protect. Strong need and narrative over a weak reality is a broken promise, the most dangerous on the list that no dashboard ever catches: the promise still lands on the surface while the delivery fails underneath, teaching every believer not to trust you next time. Strong need and reality with no narrative is a silent strength, value handed over without credit. And so on around the grid. Mend the delivery, tell the story, drop the claim, go find the need. Stop guessing what is wrong and start knowing where to aim.

None of this measures less – it just counts the thing the convenient number was always standing in for. Put the need back beside the narrative and the reality, give the intangible its full weight, and the dashboard returns to what it was meant to be: a proxy, and not the whole point.

Your intangibles were never actually unmeasurable – they were just unmeasured. That is a decision, made fresh every time an organization reaches for the nearest number because it is … near.

Hi, I’m Oscar - Founding designer at momondo, I’ve won a Material Design Award for Innovation, and I help design leaders succeed.

Subscribe now

The meditation app you bought for your design team is doing nothing, and neither is the wellness room or the breathing exercises someone scheduled for Thursday. Your designers haven’t forgotten how to breathe. What’s wearing them down is the system you built around them, and it would wear down a machine.

I’ve watched this rerun more times than I want to count: a strong designer starts fraying, leadership responds with time off and a mindfulness perk, and the designer goes back to a workflow where twelve stakeholders sign off on a button color and every deadline was invented to feel like an emergency. The perk is an apology for the thing nobody wants to change.

Why? Because we’re debugging the wrong layer. When a checkout flow leaks users you don’t tell the users to try harder, you find the broken step and you fix it. Burnout is a broken step in that flow, and we’ve been shipping the same defect for years while blaming whoever happened to hit it.

What’s actually draining them.

First up, and most common, is what happens when you strip a designer of control. Someone spends three weeks on a solution and then routes it through product, engineering, marketing, sales and whoever the CEO trusts that month, and each round of feedback contradicts the one before it. What ships is a compromise nobody believes in, assembled by committee. Do that often enough and people stop trying, because they’ve learned their effort changes nothing, which psychologists call learned helplessness and you’ll recognize as the designer who used to argue and now just asks what you want.

The counterintuitive bit is that autonomy matters a lot more than hours do. Designers who own their decisions often work longer weeks and report less exhaustion than designers on tidy schedules under a manager who owns everything, which is why a startup team can pull all-nighters through a launch and stay fierce about it. The work is theirs. Meanwhile a designer at a Fortune 500 clocks reasonable hours and updates their portfolio, because changing a hex code takes three approvals. The leash did that, and not the workload.

Then there’s the interruptions. Every Slack ping, every quick question, every meeting that should have been a message pulls someone out of the only state where good design happens, and getting back in costs minutes each time in a way that compounds all afternoon. One design team I know tracked it for a week and found the average person bouncing across fourteen different projects a day, none of it real collaboration – just drive-by requests and fires somebody manufactured that morning. By three o’clock they weren’t tired from working hard. They were tired from never getting to work at all.

Decision rights are usually a mess in both directions, so designers either drown in trivial calls nobody senior wants to make, or they can’t decide anything without escalating four levels up to someone who answers email monthly. Both burn people out, and the leaders hoarding those decisions are burning out too, buried under choices that should have died lower in the org.

The quietest one is the feedback vacuum, though. A designer ships something they spent six months on and then hears nothing at all about what happened to it, whether users liked it, whether it even launched, or whether someone killed it in a meeting they weren’t in. That silence is linked to burnout through what researchers call effort-reward imbalance, where you keep paying in and the reward never arrives. Most designers have better visibility into the office coffee supply than into the fate of their own work. People whose work has visible meaning are around four times more likely to be engaged, and you’re starving your team of the one thing that would keep them.

The myths that keep you off the hook.

The comforting story is that this is a work-life balance problem, solved with boundaries and a bit more time off, and Gallup’s data says otherwise. Engaged employees with autonomy work longer and burn out less than disengaged employees on standard hours. The way those hours get managed decides the outcome, and the count of them barely registers. Game teams grind through brutal crunch and stay energized because they’re building something they respect with people they trust, while the same hours in an agency where every idea dies in committee produce collapse, boundaries or no boundaries. Bezos made this point years ago when he threw out the word “balance” for something closer to a circle, where the two sides feed each other instead of trading off.

The second story blames the person, because “that designer couldn’t hack it” is an easier sentence to say than “our approval process would break anyone.” The evidence doesn’t support it. Gallup’s strongest predictors of burnout are all organizational, starting with unfair treatment and running through unmanageable workload, unclear roles, absent managers and time pressure nobody could meet. The WHO defines burnout as chronic workplace stress that hasn’t been managed, and it drew that definition specifically to pull the thing out of the category of personal weakness. Employees who feel treated unfairly are 2.3 times more likely to hit severe burnout, and no amount of resilience training survives contact with a toxic manager or a team that’s been understaffed for a year.

And then the vacation, which won’t hold. Your burnt-out designer takes two weeks, comes back lighter, and within days is back in the same soup of approvals and interruptions that drained them in the first place. Time off pauses it, that’s all. If anything, a rested brain works out faster that the job is the problem.

So what do we change?

Start with the approval map. Write down every sign-off your designers need before something ships, and you’ll find something absurd in there. I once found a team that needed seven approvals to change error-message copy and one to spend fifty thousand dollars on stock photos. Then give designers decision rights over their own domain, and not the right to recommend. Start small if you have to, let them own typography without a committee, and watch what a month of that does to their energy. When a European tech company hit a burnout wall, its new VP of Design skipped the resilience training and ran workshops where designers mapped their own pain points instead, and they named the approval layers and the unclear priorities themselves. Leadership collapsed the sign-offs and started ranking projects quarterly. Turnover dropped over the following year, and self-reported burnout fell with it. People defend what they helped build.

Then protect focus time like it’s revenue, because it is. At KAYAK our No Meeting Fridays are an admission that deep work needs defending and won’t survive a calendar chopped into thirty-minute fragments. Guard the mornings for the hard creative work while brains are fresh, and push collaboration and feedback into the afternoon slump where they belong anyway. Google’s reliability teams cap manual toil at half of an engineer’s time and keep the rest for work that compounds, and it cut their turnover. Then track focus time the way you track velocity, since the thing you don’t measure is the thing you let rot.

Close the feedback loop after that, and keep it closed. Everything a designer ships should come back to them with what happened to it, close to when it happens rather than buried in an annual review. One team I worked with ran a live dashboard of user reactions to recent launches, and another sent a monthly one-pager tying design calls to actual metrics. When designers can see impact, even a failure turns into something they learn from instead of a void they shouted into.

Finally, hunt the waste upstream. Map where designs get stuck and what share of your team’s hours go to rework rather than new work. Lock brand decisions before design starts and not halfway through. Pull designers into discovery instead of handing them finalized specs that bake in constraints nobody can solve later. And publish the waste numbers, every single one of them – how many design hours died on work that never shipped, how many revisions traced back to late input. You can’t fix what you refuse to count.

None of this makes your designers tougher, which was never the point. The chaos is the thing you’re supposed to design out.

So the choice is yours and not your team’s. Tomorrow you can walk in, ask what one thing they’d change if they had the power, and then actually kill it. One approval you delete, one recurring meeting you end. Or you can renew the meditation subscription and keep filing burnout under personal failure, which is cheaper, easier, and works right up until your best designer’s portfolio goes live.

Hi, I’m Oscar - Founding designer at momondo, I’ve won a Material Design Award for Innovation, and I help design leaders succeed.

Subscribe now

Every prototype your AI tools produce is a remix of something that already shipped. You keep blaming the model, the prompt, the iteration count. Blame your first sentence instead.

A rebellious Chinese painter-monk diagnosed this failure in the late 1600s, and a Wharton researcher confirmed it with data some three and a half centuries later. Between them, they explain why your AI-assisted work feels derivative and what to do about it.

The problem with copying old masters

Shí Tāo painted in a period when Chinese art had calcified around the “Four Wangs,” a group of painters who had turned excellence into imitation. You studied the old masters, you reproduced their strokes, you got praised for fidelity. Creativity had become draw-by-numbers with better calligraphy.

Shí Tāo watched this and called it what it was: creative death by a thousand brushstrokes.

His alternative centered on what he called the primal mark. The first stroke on the paper doesn’t just influence the painting. It shapes everything that follows. Commit to a conventional opening stroke and the painting is conventional before you’ve mixed your second ink. The rest is elaboration.

For over 300 years this stayed philosophy – then someone tested it.

Science catches up with the monk

In 2014, Justin Berg at Wharton ran four experiments with around 800 participants. He gave people different starting points for creative tasks and measured the novelty and usefulness of what they produced. The design was sharper than “do starting points matter.” He tested whether different types of starting points led to predictably different outcomes.

They did. People who started from familiar concepts produced work that was useful and predictable. People who started from purely novel concepts produced work that was original and often useless. And people who started from what Berg called “integrative” concepts, blends of familiar and new, produced ideas that scored high on both. They got the breakthrough and the utility.

The worst bit is how much the starting point locked in. Once you commit to developing an idea from a particular origin, Berg writes, “the fate of any ideas that grow from it may be largely sealed.” We agonize over final details and spend almost nothing on the opening move. That’s obsessing over the steering wheel while someone else picks the road.

You can’t iterate your way out of a bad first stroke.

Your brain is a prediction engine that never switches off

One more piece makes this click. Psychologists call it “nexting,” and your brain is doing it right now. It predicts the next word in this sentence, the next stair under your foot, the arc of a frisbee, what follows “It was a dark and stormy...” Hundreds of forecasts per second, all outside awareness. When a prediction fails, you feel a jolt and your slow, deliberate thinking wakes up to investigate. The rest of the time, prediction runs the show. It’s why you can walk without computing each step and read without sounding out letters.

Large language models are nexting machines by construction. Trained on enormous datasets to predict the next token, they are prediction engines wearing a chat interface. When you ask Figma Make, or Claude Design, to design something, it forecasts the most probable design given everything it has seen.

Two prediction engines, one human and one artificial, are now sitting down to make things together. You can guess what happens next. So can they.

The AI orthodoxy trap

AI has become the new orthodox school. Type “Create a dashboard for travel management” and you’ve asked a prediction engine to forecast what follows those words based on every travel dashboard in its training data. What comes back is the design equivalent of a Four Wangs painting: competent, recognizable, spiritually empty.

Berg’s framework names the mechanism. That prompt sets a familiar primal mark, one built from conventional ideas within the domain. His experiments show this anchors you toward usefulness at the expense of novelty. You’ll get something that works. You won’t get anything new.

The AI isn’t lazy. It’s doing its job, predicting its way through the most probable sequence of design decisions. You’ve asked for the statistical average of all travel dashboards in slightly different visual clothing.

Your own brain makes it worse. You see the first outputs, your prediction engine locks on, and now you’re collaboratively predicting your way toward the most probable solution. Human and machine reinforce each other’s pull toward the familiar. A feedback loop of two nexting engines, each confirming the other’s guesses.

No wonder everything feels derivative.

The anchor that seals your fate

Berg drew on the psychology of anchoring to explain why first moves dominate. “The initial content in the primal mark may impact novelty and usefulness disproportionately more than content added later in creative tasks,” he writes. Start from a familiar mark and “familiar schemas dominate their thinking,” crowding out whatever novel associations might have surfaced.

This is why those long polishing sessions with AI feel so futile. You refine and refine and never escape the gravitational pull of the first conventional idea. Berg found that novelty is “more rigidly anchored by the primal mark than usefulness.” A familiar starting point puts a ceiling on how original the result can ever be, and no amount of iteration raises it.

Shí Tāo saw this intuitively when he described painters “beclouded by things,” dragged down by “a thing’s dust.” Our dust is the sediment of every interface ever shipped, compressed into training weights and reinforced by our own predictive habits.

The method of no-method

Berg’s data points at integrative primal marks: starting points that fuse familiar and new content. Building one requires analogical thinking, searching for “higher-level, abstract parallels between two or more ideas” rather than surface similarities.

Shí Tāo called this the method of no-method, which sounds like wordplay until you apply it. It means freedom from dependence on predictive patterns. You notice when you’re predicting and you choose to break the chain.

Applied to AI, the shift looks like this. Instead of “Create a dashboard for travel management,” you open with:

“Explore how ancient migration patterns and wayfinding rituals could inform how modern travelers move through their journeys. Consider the intuitive connection nomads had with landscapes and the data streams of contemporary travel. What emerges when we honor both the need for discovery and the reality of complex logistics?”

Now you’re asking the machine to synthesize something that doesn’t exist in its training data as a settled pattern. You’ve forced it out of prediction mode. You’ve set an integrative primal mark.

And the marks pay double. Berg’s experiments showed integrative starting points beat familiar ones on novelty and beat purely novel ones on usefulness. The mechanism runs through the analogical thinking they force. Hunting for abstract parallels between distant ideas surfaces “more fundamental — and thus more novel — ways of recombining” them, while the familiar elements “enhance the clarity, meaning, legitimacy” of the result. The familiar half helps people understand the novel half. The novel half keeps the familiar half from going stale.

Berg’s own examples of the three primal marks in action. NGL, I want that recording shoe. 🙃

The painter moves the ink

“The painter moves the ink, the ink does not move the painter.”
— Shí Tāo

Mastery means keeping creative agency while using tools, rather than serving your tools’ tendencies. AI generates results so fast and so confidently that it’s easy to slide into outsourcing your creative decisions to statistical patterns without noticing the handover.

You see the pattern everywhere. Designers take AI wireframes as starting points and polish them. Writers take AI drafts and edit them. Strategists take AI frameworks and customize them. None of this is wrong. All of it is capped. The AI’s prediction has become your primal mark, so you’re anchored to its statistical read of the problem rather than your human read of what’s needed.

Berg’s research suggests the inversion: use AI as raw material for building integrative primal marks, never as a source of familiar ones. The ink, in other words. You hold the brush.

Three stages for breaking the chain

Stage one: interrupt your own prediction. Before you open any tool, catch your brain mid-forecast. Skip “what should the solution look like” and ask instead: what assumptions am I carrying? What would the obvious path produce? What human need sits under the stated requirements that could support a different starting point? This is strategic thinking about your primal mark before you commit to one, because once either engine starts predicting, Berg’s data says the path is hard to leave.

Stage two: prompt to force synthesis. Craft prompts that push the AI outside its probable chains. Pair domain requirements with a perspective from somewhere else entirely. Ask it to hold contradictions open rather than resolve them. Use it to question the problem definition, not just answer the problem as stated. You want a starting point that makes you slightly uncomfortable. Discomfort is evidence you’ve left familiar territory.

Stage three: watch for relapse. As you iterate, stay alert for the moment outputs start feeling “right” too fast. That ease usually means you’ve slipped back into collaborative prediction with the machine. Berg found “usefulness is more flexible than novelty”: you can make a strange idea practical later, but you can’t make a familiar idea original later. When in doubt, err toward strangeness. Surprise is your compass.

The mirror nobody ordered

One more opportunity hides in all this. An AI trained on human cultural output is a mirror reflecting the collective creative patterns of our civilization. When it generates something predictable, it’s showing you the statistical average of human work in that domain. When it generates something odd, it may be surfacing combinations that exist in the data but that nobody has consciously noticed.

Berg found people gravitate toward familiar primal marks 60 to 80 percent of the time. The bias toward the safe and conventional is in us, not just in the model. The machine lets you watch that bias operate in real time and work against it on purpose, pairing the collective intelligence baked into culture with the one capacity prediction can’t reach: imagining futures that have never appeared in any dataset.

Your next project

Pause before you open the tool. Ask whether the mark you’re about to set leads anywhere but the statistical average. Both you and the AI are prediction engines. The creative opportunity lives in breaking those chains deliberately and using the machine’s pattern-matching to reach territory neither of you would find alone.

When outputs feel right too quickly, suspect the trap. When you hit something original, trace how it happened. You’ll usually find an integrative mark at the origin, a starting point that honored both a human need and a possibility nobody had trained on yet.

Culture can keep predicting its way toward more of the same, or it can imagine something it has never seen. Your first stroke decides which one you’re contributing to.

Hi, I’m Oscar - Founding designer at momondo, I’ve won a Material Design Award for Innovation, and I help design leaders succeed.

Subscribe now

Everything you’ve read about change management warns you against paralysis. Paralysis would at least be visible. What really happens is that engineering doubles down on velocity, design disappears into pixel-perfect mockups, product turns into a feature factory, and every dashboard glows green while the organization drifts. The work keeps happening. Nobody knows where it’s leading.

I’ve watched this turn brilliant design leaders into micromanagers inside a month. One week they’re setting product vision, the next they’re relitigating whether a loading animation should run 200 or 300 milliseconds. Change didn’t make them worse leaders. It revealed that their strategic thinking only worked when someone above them kept the sky clear.

Busyness is a stress response

The execution trap is the false comfort of “at least we’re shipping something” that kicks in when the bigger picture goes cloudy. It feels rational. It’s a defense mechanism.

The clearest evidence comes from a place with much higher stakes than a product org. In August 1949, fifteen smokejumpers parachuted into Mann Gulch, Montana, to contain what looked like a routine fire. When it blew up and started chasing them uphill, their foreman, Wagner Dodge, did something unprecedented: he lit a fire in front of himself, stepped into the burned patch, and shouted at his crew to drop their tools and join him. They didn’t. They ran, carrying their saws and shovels up a 76-degree slope, and thirteen of them died. Some were found still wearing their packs.

Organizational psychologist Karl Weick spent years on why trained professionals held onto twenty pounds of steel while running from a fire. His answer: under acute uncertainty, people cling to whatever made them competent before, because dropping it means admitting the situation has changed beyond their existing skills. A firefighter without tools is just a person running. A designer who stops polishing mockups mid-reorg is just a person who doesn’t know what the company is doing.

Weick called the underlying mechanism cognitive narrowing. Threat shrinks perception down to the immediate and familiar. Your team isn’t choosing to ignore the bigger picture during change. They can’t see it. The narrowing has already happened by the time you notice the symptoms: velocity worshipped as a north star, “we completed 15 story points” replacing “we moved retention,” leaders retreating into process because process is the one thing still in their control.

Everyone is moving fast. Nobody is steering.

The three altitudes

To see the way out, map where leaders actually operate. There are three altitudes.

Ground level is the 50-foot view. Sprint planning, task management, today’s fire. Time horizon of days to weeks. Its failure mode is executing blindly without asking whether the work still fits anything.

Sky level is the 50,000-foot view. Market position, long-term vision, resource allocation. Quarters to years. Its failure mode is abstraction so thin the ground never feels it.

Eye level sits between them, around 5,000 feet. This is where sprint work gets connected to broader themes, where fuzzy strategy gets translated into coordinated action. Weeks to quarters.

During organizational change, the sky clouds over. Strategy is being renegotiated somewhere above you, and no amount of town halls will un-cloud it, because the honest answer is that leadership doesn’t know yet either. This is why the standard advice for change, communicate the vision harder, fails so reliably. You cannot communicate a vision that is still being fought over. Teams smell the uncertainty under the confident slides and narrow further.

When the sky is unreadable, ground-level teams look at their feet. The only leaders who can help are the ones operating at eye level. And eye level is precisely the altitude most leaders abandon first, because it’s the hardest to defend in a status report.

Direction without a destination

Eye-level leadership does something neither strategy nor execution can: it provides direction when the destination is unclear.

Strategy researcher Richard Rumelt calls the tool a proximate objective, a near-term goal close enough that people can actually see how to hit it. When distant goals turn to fog, the proximate objective is what keeps a group coherent. The eye-level move sounds like this: “I don’t know where the company’s AI strategy lands. I do know that for the next quarter we’re fixing data quality in existing features, because that pays off in every version of the future.”

Notice the structure. Not a vision. Not a task. An approach.

I saw this work on an team mid-restructure. Company strategy was officially “under review,” which everyone correctly read as “being fought over by two executives.” The team was rebuilding a booking flow, and at ground level that’s all it was, a feature on a list. Their lead reframed it as an approach: we reduce anxiety during booking, we show real-time availability, we surface fees before the payment step even when that costs us upsell conversion. That single sentence, does this reduce anxiety or create it, then settled hundreds of micro-decisions over the following months without a single escalation. The strategy stayed foggy the entire time. The team never felt lost.

That’s the test of eye-level work. When your team can articulate their approach, they can make decisions faster than you could ever make them for them. When they can only articulate their tasks, every ambiguity travels up to you, and you drown in exactly the micromanagement you swore you’d never do.

So here’s what you do next

You don’t fix cognitive narrowing with an offsite. You fix it by changing the questions asked in meetings that already exist.

Swap “what did you ship this week” for “what are you learning about our approach.” The first question rewards motion. The second forces someone to lift their head, even for thirty seconds, and check the horizon. Do it every week and the team starts preparing answers at eye level, which means they’ve started thinking there.

A design director I worked alongside ran a monthly session she called connect the dots. No agenda, no deliverables, one prompt: what story is our work telling? The first session was awkward silence. By the third, her team was catching things like “we’ve solved onboarding friction the same way in three unrelated features, that’s not coincidence, that’s an unstated principle.” Patterns that sprint tunnel vision had made invisible. The session cost an hour a month and replaced most of what her old strategy reviews pretended to do.

And when someone says “I think we’re heading the wrong direction,” treat it as the system working. Amy Edmondson’s research on psychological safety keeps landing on the same point: candor is a performance feature, not a comfort feature. In stable times a team can survive without it. During change, a team that can’t say “this feels wrong” out loud will express the same doubt through slowdown, attrition and quiet sandbagging instead. The doubt doesn’t disappear. It just stops being useful to you.

None of this requires the strategy to resolve. That’s the point. Waiting for clarity before providing direction gets the sequence backwards. Direction is what teams need most in exactly the periods when clarity is unavailable.

The smokejumpers who died at Mann Gulch were fit, trained and fast. Speed wasn’t their problem. They were sprinting with full competence in a direction that no longer made sense, holding tools for a job that no longer existed. Wagner Dodge survived because he stopped, read the situation at a different altitude, and did something that looked insane at ground level.

The next reorg will hand you the same choice in a milder form. You’ll either give your team an approach, or you’ll give them a velocity target. Only one of those is leadership.

Stop waiting for strategy to get clearer. Start building the muscle that lets you provide direction even when the destination is still coming into focus.

While the last stragglers amongst designers still debate how AI might “augment” their process, your product manager two desks over is sketching wireframes over coffee, and your CEO who still asks the assistant to print their email is generating a passable prototype by typing a sentence into a box.

The thousand-intern army that AI is, can execute any direction you give it. Its own idea of creative inspiration is “what if this button were slightly bluer.” It’ll spin out a thousand variations of a layout until the moment you ask which one is better and why, and it will likely give you a very confidently bad answer.

So your value designer starts to move to the instinct that picks the one good option out of the thousand, and to something larger that we haven’t really named yet. Because this breaks the assumption underneath everything we do – that people move by hand through windows, icons, menus and pointers.

The WIMP model that built the profession is dissolving fast.

A comfortable forty-year run.

In 1981, Xerox unveiled the Star Information System and freed us from remembering commands that read more like incantations written by wizards. Suddenly your grandma could use a computer without a computer science degree. And the desktop metaphor they bundled everything into was so obviously right that we’ve never really questioned it since.

So for four decades, designers built whole careers becoming virtuosos of this one visual approach. We made windows feel like windows, complete with the satisfying snap when they lined up. We turned abstract functions into little icons your mom could read, and choreographed menus that revealed exactly what people needed exactly when they needed it.

As UX guru Jakob Nielsen points out, nearly every interface built after time follows the same WIMP style. We didn’t just adopt a pattern that worked – we internalized it until questioning it felt almost unthinkable, and got so good at working inside the limits that we stopped seeing them as limits entirely

Cracks become chasms.

WIMP was built for one reality: one person, one keyboard, one mouse, one fixed screen. It’s perfect for that and increasingly awkward for everything else. Voice was the first departure and hardly the apocalypse everyone predicted, because talking to your smart speaker is fine and debugging a stubborn CSS layout out loud is not. (Unless you’re one of those Wispr Flow freaks.)

We already have about ten million people walking around with the answer strapped to their face like the Ray-Ban Meta glasses. You ask them something and the answer arrives in your ear or across a lens. No window to open, no icon to tap, no menu to walk, no pointer to aim. The entire WIMP vocabulary is just absent, and nobody wearing them notices it’s gone. The Xerox Star team would either call it magic or assume someone had spiked their coffee.

Software like Siemens’s DesignCenter now watches which tools and commands you reach for, and in what order, and then reshuffles its own menus and toolbars to surface the ones you’re about to need. Netflix and Amazon have been doing a cruder version of this to their front doors for years. The interface simply reorganizes itself around you instead of waiting for you to hunt through it, which means interfaces are starting to have opinions about what you’ll do next.

AI as the interface designer.

And in the same way adaptive interfaces started taking control away from the user, AI is now taking over the design itself. It’s doing the interface design now, generating what gets rendered, per person, per moment.

Nielsen again describes generative UI as interfaces generated in real time to fit one user’s needs and context, and uses the example of an airline app that rebuilds itself for a frequent flyer with dyslexia: dyslexia-friendly type, her home airport already filled in, the red-eyes she never books hidden. That interface exists for exactly one person. Old GUI design was building a stage set, fixed for every performance. This is closer to directing improv.

Everyone’s answer to this is to point AI at the design system and let it assemble. Connect the component library, expose it over an MCP server, and the machine has everything it needs. Most teams racing to get AI-ready are doing exactly this and calling it done.

It doesn’t work though, and we know it doesn’t because it’s been tried already. One example from Yesenia Perez-Cruz, who led Polaris at Shopify for five years, was giving Claude Code full access to the Polaris library and its MCP server, every component and composition pattern the system had ever documented. Then she asked for a UI to confirm changes to a discount code. She got the resource detail layout with some styling to mark what had changed. Then she asked for a UI to extend the expiry date across thirty discount codes. She got the same resource detail layout, with a table dropped inside one of the cards.

Those are two completely different jobs, one person reviewing a single change and another editing thirty records at once, and the output came back nearly identical both times. The association between “discount” and “a page that looks like this” was strong enough that she couldn’t shake it loose, because the library told the model what a discount looks like and never told it what a discount is for.

The way out of the “more of the same” maze, is adding a semantic layer to the mix, rather than more components. Simple markdown describing the objects themselves, the surfaces they can appear on, and the signals that indicate what someone is trying to do. Now running that same model, same library, same two prompts, the confirmation came back as a compact modal with the changed values marked and an impact summary, because reviewing one edit doesn’t need a full page. The bulk request came back as a proper batch editing screen with the new end date called out in its own column, the density adjusted to the task.

And then something new: the AI composed a bulk date picker that didn’t exist anywhere in the library, reasoning that if you’re extending thirty dates you probably want to select them all and set the value once. The system now going beyond just re-using what had already been defined, and doing so with reason.

Even the way it narrated its own work changed. Without the context layer it listed what it had picked: a banner, an index table, checkboxes, a sidebar, status badges. With the context layer it named the intent, named the surface, and explained the decision.

Your component library documents appearance, which is the part the industry keeps missing. It says nothing about when a thing should be used, what the person is trying to accomplish, or which presentation serves that goal, because for forty years the designer supplied all of that from their own head at the moment of use. It never needed writing down. Hand the library to a machine and it fills the gap the only way it can, by reaching for the most statistically common arrangement and calling it a decision.

Feed the same machine your intent and it starts making choices you’d defend in a review.

“The uncomfortable truth is,”

Yes, seriously. Let’s be honest about what most of us do all day: we apply templates within constraints. We arrange existing elements along established patterns and dream of real creative work. We’re masters of variation inside fixed spaces, and none of that is a small skill, because there’s craft in making an interface feel effortless.

But when AI throws out thirty variations in the time it takes Figma to load, the market value of that pixel-perfect refinement collapses. Churning out asset variations, producing responsive layouts, A/B testing visual alternatives: generative tools already do all of it with inhuman speed and consistency.

“Junior designers are done. AI does thousands of icons in seconds. Only creative directors survive.” — Some design leader, probably.

The jobs that disappear are the ones defined by deliverables instead of outcomes and impact. The designer whose whole value is a beautiful static mockup is in trouble as teams move to live AI prototyping, and the UX lead who hand-documents every screen state will watch AI generate those states on the fly.

None of it is worth mourning though, because clearing out the old production roles opens room for harder work, and our new semantic layer is exactly the kind nobody currently owns.

The question now isn’t where you use AI, but where you don’t.

For a while the honest advice was that you had a few years to adapt before AI-driven design became the norm. That deadline came and went while nobody was watching.

Figma spent 2024 shipping its first AI features as opt-in extras you could ignore. By Config 2026 that framing was gone. Make turns a sentence into a working, editable prototype. An agent co-designs on the canvas. The tool checks your work for accessibility and hands structured code to engineering without being asked. None of it is a beta you switch on. It’s the floor, and clients already expect the iteration speed it buys. Apple, Google and Microsoft are baking AI-driven interface APIs into their design guidelines the way they once baked in responsive principles, so the baseline keeps rising underneath you whether or not you move.

As Bob Baxley put it (, and I’m paraphrasing here):

“AI is taking us from throwing spaghetti at the wall to see what sticks, to getting a spaghetti throwing machine that can do it much faster.”

Sure, it's an upgrade, but we're still solving the problem the wrong way, just making the mistakes faster.

Which is why the interesting question has flipped. It’s no longer whether you use AI, because everyone does, and the holdouts are already slower. The question worth your judgment now is where you deliberately don’t use it, and why. Which decisions you keep human on purpose. Where a generated interface is faster but quietly worse. Where personalization curdles into manipulation, or the efficient answer strips out the friction a person actually needed. Nobody audits a screen that only exists for one person and never renders the same way twice, which means the guarantees we used to check at the end now have to be built into the thing doing the generating. Knowing where to withhold the machine is turning into a sharper skill than knowing how to run it.

The move back to fundamentals is exactly what design needed, and plenty of designers will still refuse to make it, because arranging four elements on a screen is comfortable and owning how a system treats a million different people is not. And that’s the real split.

This article was edited July 2026 to add better examples, and now has a follow-up too:

Hi, I’m Oscar - Founding designer at momondo, I’ve won a Material Design Award for Innovation, and I help design leaders succeed.

Subscribe now

Your brain is running malware from 1984

Your brain is lying to you about AI, and it’s running on a script written before the transistor got cheap. That anxiety you feel scrolling past another LinkedIn post about designers getting replaced by ChatGPT is a decades-old horror movie playing in a part of your head you don’t have admin access to, and you’ve mistaken it for analysis.

I’d like to prove it.

Henry Modisett, Founding Designer at Perplexity, made an interesting point: almost every AI story we grew up on ends the same way. Terminator, Blade Runner, I Robot. The machines wake up and we lose. Most were written long before anything resembling a modern model existed, and they became the default operating system we run to interpret every demo, every jaw-dropping release. A language model writes a passable brief and some ancient subroutine fires: Skynet is loading.

The fear tracks Hollywood’s AI, not the one shipping in your tools.

Harvard psychologist Daniel Gilbert spent a career on the mechanism underneath this. When we hit an ambiguous situation (and AI is nothing but ambiguity right now), the brain doesn’t weigh the evidence neutrally. It fills the gaps with whatever story feels most familiar, then hands you the result as if it were a fact you observed rather than a guess you manufactured.

The thing you’re actually scared of isn’t replacement

The frightening part is the not-knowing. You can’t predict whether AI replaces designers, how, when, or how far it goes. Certainty you could plan around. This open-ended blank is what’s eating you.

That uncertainty trips what Gilbert calls the psychological immune system, the mental defense that shields you from existential threats and, like any immune system, sometimes attacks healthy tissue because it can’t tell the difference. Yours has decided a text-generation tool is a pathogen.

So it does what immune systems do. It cooks the facts. You scrutinise every AI failure like a forensic examiner and wave through every success on a glance. You take today’s clumsiness and project it onto next year’s model, assuming the collaboration will always feel as foreign as it does this afternoon. What your defenses can’t model is how fast you’ll adapt once the thing is just part of the furniture.

We have done this three times already

Desktop publishing was going to kill graphic designers. The web was going to make print skills worthless. Mobile was supposed to finish off web designers for good.

None of it happened. The tools changed, the job absorbed them, and the people who’d built their identity around a specific piece of software had a rough couple of years while everyone else got on with it. This moment feels different only because you’re standing inside the transition instead of reading about it later.

Step back and look at what AI does in design work today, in 2026, stripped of the soundtrack. It iterates faster than you can. It generates a hundred starting points you’d never have sketched. It swallows the production grind that used to eat the hours you’d rather spend on judgment. Agents now act on your behalf, wiring up flows and filling in states while you’re looking elsewhere.

Which makes it a sophisticated Photoshop that talks back.

The generative UI shift already underway changes what the work is: from pushing pixels toward directing outcomes, from shipping static screens toward shaping systems that assemble themselves in context. That’s a bigger job.

Your defenses are writing your career strategy, and they’re bad at it

Unconscious bias steers your decisions whether you sign off on it or not. Hand your AI strategy to Hollywood reflexes and an over-caffeinated immune response, and you’ve put the most primitive part of your brain in charge of your next five years.

Run the exercise. Assume AI becomes as fundamental to design as the computer became in the 1990s. Now picture the designer in 2026 who refuses to touch it, on the same trajectory as the one who insisted on hand-lettering everything in 2010. Competent, principled, and quietly becoming unhireable.

Your defenses will offer you respectable-sounding cover. AI lacks real creativity. Clients pay for the human touch. It’s a bubble. That machinery is superb at manufacturing dignified reasons to avoid a thing you’re afraid of. It’s the same machinery that helps people make peace with situations they’ve decided they can’t change, which is useful for grief and useless for strategy.

The designers who come out of this ahead won’t be the ones who held out longest. They’ll be the ones who got their hands dirty early enough to shape how the tools fold into human judgment, before the shape got decided without them.

So pick. Either you name the reflex for what it is and choose deliberately, or you let a threat-detection system built for saber-tooth tigers allocate your career. There’s no third option where you keep the anxiety and the outcome you want.

Your skills (reading ambiguity, understanding what people need before they can say it, making creative leaps with half the information) are exactly what this era runs on. None of that is at risk. What’s at risk is whether you point those skills at the new tools instead of against them.

Hi, I’m Oscar - Founding designer at momondo, I’ve won a Material Design Award for Innovation, and I help design leaders succeed.

Subscribe now

You're interviewing a designer who immediately understands your product, speaks your industry jargon fluently, and nods knowingly at your biggest challenges. Perfect fit, right? You hire them, and six months later they're producing the same solutions your team has been cycling through for years.

Meanwhile, the candidate who asked "Why would users want to do that?" gets passed over for not understanding the business. Except they were the only one actually thinking about the users.

This is the fallacy of the plug-and-play designer. We mistake familiarity for insight, domain knowledge for fresh thinking. Teams get really good at solving yesterday's problems with yesterday's solutions, while billion-dollar opportunities hide in the questions nobody thinks to ask.

Fresh eyes see what experience misses.

Sometimes the most expensive question you never ask is "Why are we trying to make a better striped soap?"

In the 1970s, Procter & Gamble spent months trying to compete with Irish Spring's green stripe. They brought in a young designer named Min Basadur who asked that seemingly obvious question and completely reframed the challenge from "How might we make a better green stripe bar?" to "How might we make a more refreshing soap?"

Three hours later, they had Coast soap. One question. Billion-dollar product.

The intern who asks "Why do we have 17 stakeholders in this meeting?" isn't being difficult. They're spotting dysfunction everyone else learned to work around. The new hire questioning why legal reviews design mocks might be uncovering a process problem adding months to every project.

"What if people could rent out their spare rooms to travelers?" — Two cash-strapped designers in San Francisco, 2007

That question sounded ridiculous to hotel industry experts. Those designers were Brian Chesky and Joe Gebbia. Their "dumb" question became Airbnb.

The difference between productive naivety and willful ignorance is curiosity versus ego. Fresh eyes ask "Help me understand the constraint here." Troublemakers ask "Why?" just to watch things burn.

We systematically train fresh perspective out of new hires. They ask "Why do we do it this way?" and get hit with "That's just how we do things here" (the six words that kill innovation, according to Steve Blank). After a few rounds, they stop asking. Often nobody remembers why the process exists in the first place.

How to spot the good question-askers.

I've got a go-to interview move that reveals everything about how someone thinks. I show candidates a polished piece of our design work — maybe our booking flow or onboarding sequence — and ask them to walk me through what they notice.

The weak candidates ask surface-level questions about visual details. "What font is this? Why did you choose this shade of orange? Do you use relative nested corner radii?" They're focused on the what and how of design decisions. (OK, that last one about corner radii is a green flag, not gonna lie.)

But the naturally curious ones? They squint at the screen. They ask follow-up questions. "This is really clean, but I'm curious about the mobile breakpoint here. How do users typically navigate between these sections?" They're not trying to flatter me. They're trying to understand the problem we were solving.

Here's where I test their intellectual humility. I'll point out a potential improvement in their portfolio work — nothing harsh, just a thoughtful observation about a different approach they could have taken.

The candidates who freeze up or immediately justify their choices are telling me everything I need to know. "Well, we couldn't do that because of technical constraints" or "The stakeholders wouldn't have gone for it." They're in defense mode, protecting their ego instead of engaging with the idea.

The ones I want to hire get visibly excited. "Oh, that's an interesting note… I never thought about approaching it that way. That could have solved the accessibility issue we ran into later." They treat feedback like a puzzle piece they've been looking for, not an attack they need to repel.

The best candidates balance respect for requirements with healthy skepticism. They prioritize understanding the real problem before solving it.

Designers who ask clarifying questions during interviews become dramatically better at stakeholder management. Picture an airline's customer experience team being told to "make the boarding process feel more premium." The person who asks "What do you mean by premium feel?" instead of nodding and guessing always produces work that matches what stakeholders actually had in mind.

Building teams that make your whole organization smarter.

Your organization will immediately try to train curiosity out of new hires. They'll ask "Why does this process exist?" and get shut down with "Because that's our process."

Your job as a leader is creating psychological safety for questioning while maintaining execution focus. It won’t always be easy.

Implementation strategies that work for me:

• Explicitly tell new hires their fresh perspective is valuable. I literally say "Your job for the first 90 days is to ask dumb questions. We've gone blind to our own problems."

• Pair curious newcomers with experienced designers who welcome questions, not those who get territorial about their expertise.

• In team meetings, praise good questions as much as good answers. When someone asks "Wait, why are we solving this problem again?" dig in instead of rushing past it.

• Protect questioners from institutional antibodies. When senior stakeholders grumble about the new person "not understanding how we work," that's usually a sign they're onto something important.

Experience tells you which questions have been asked before. Curiosity tells you which new questions haven't been asked yet. You need both on a team, but if forced to choose, I'll take the curious novice over the complacent veteran every time.

The paradox of design expertise is that the more experienced you become, the harder it gets to see your work through fresh eyes. Veteran designers have to actively fight against their accumulated assumptions and biases. Sometimes the most efficient way to maintain that beginner's perspective is to hire people who actually have it.

Curious designers transform entire organizations. Someone questions why user research happens after designs are finalized, and suddenly the whole product development process gets smarter. Someone asks "Why don't we ever talk to users who didn't end up booking?" and the entire growth strategy shifts.

"Experience tells you which questions have been asked before. Curiosity tells you which new questions haven't been asked yet."

The compound interest of intellectual curiosity.

The most dangerous phrase in design isn't "That's how we've always done it." It's "Users love this feature." The first acknowledges tradition exists. The second wraps ego in the flag of user research.

Teams that celebrate dumb questions build organizations that can adapt, learn, and spot opportunities experts miss. They create cultures where a 22-year-old intern can save months of wasted process, where a fresh hire can prevent expensive duplicate features, where someone can ask "What if users don't actually want this?" without getting fired.

The designers who ask obvious questions aren't slowing you down. They're the canaries in your coal mine, spotting expensive mistakes hiding in plain sight. They're building teams that make your whole company smarter, one beautifully naive question at a time.

When a candidate asks "Why would users want to do that?" instead of immediately nodding along with your product vision, don't pass them over for not understanding the business. Thank them for being the only person in the room actually thinking about the users.

We all agree creativity can’t be rushed, that good work needs its iterations and that inspiration lands when it lands. And design managers have become the strongest defenders of that notion. So you explain why research takes three weeks, you defend your second round of iteration, you educate stakeholders about a creative process that resists deadlines, and every one of those explanations confirms the very thing you’re arguing against. The room hears that design is the part that can’t keep up.

And in a raw comparison it looks bad. Engineering ships code in hours while a designer spends two days on visual hierarchy. Product decides something in a Tuesday meeting while design goes off to “explore the problem space”, and anyone watching the calendar draws the obvious conclusion.

Agile was tuned for engineering and runs beautifully there. Tickets, story points, stand-ups, a burndown chart that meticulously shows progress and delivery. And then design work goes into that same machine, asking it to declare up front how long it will take to make the experience better, or to work out why conversion dropped, without spending too much time thinking about it. Neither of those is a story, or can be estimated effortlessly in Fibonacci series. And what comes out the other end instead measures how badly the container fits, and says nothing at all about how fast your designers work.

Companies that ship fast instead get there by pulling friction out of everything sitting around design, not by hiring “quicker” designers.

Find your bottlenecks before you defend anyone

Your first instinct as a design manager when someone calls the team slow is to protect them, and that instinct puts you in the weakest position in the room - explaining instead of diagnosing. Someone just handed you a systems problem with a people label. Take off the label before you start arguing.

Start by counting. How many engineers do we have on the product team versus how many designers. Most organizations land somewhere around six or eight engineers to a single designer, and that designer is often split across two or three initiatives while also being asked to run research, make wireframes, design the UI, write the UX copy and also keeping three different sets of stakeholders informed. Nobody is slow in that arrangement. The math just doesn’t math.

Then you look at when design actually gets invited. What often happens in tech is that requirements are already locked by commercial, the timeline is set by the PM, the architecture is agreed with tech, and somewhere after all of that someone remembers to add design to the mix. And all you’re left with now is repair work on decisions nobody made with a user in the room – and repair always takes longer than the original would have.

The final part is the scope. Monday it’s a settings page. Wednesday it needs enterprise permissions, three API integrations and could it work on my iPad too? Each of those changes fans out into a week of revisions, and when the deadline slips the slip often gets marked against design.

Buy speed with better choices

Once you’ve snapped out of your defensive mode, you can start deciding, and the first decision is where quality pays.

A small handful of design decisions carry most of the impact for actual users, so spend there and stop pretending the rest deserves equal care. A Kano diagram always comes in handy up front. Look at the core flows, the interactions people repeat daily and the moments that shape how the brand feels all earn the extra week. Internal admin screens and the edge case that fires twice a month aren’t in quite the same category.

And sometimes shipping something rough is the right call to make. UX debt behaves like technical debt, fine to borrow against and darn dangerous to forget. So while you’re shipping the fix that stops the leak, you also need to book in the thing that isn’t just a fix.

Design systems are where this really compounds. A shared system lets you build better and faster at the same time, better because people understand a consistent experience more easily and faster because everyone works from the same vocabulary. And a smart and structured approach does more for velocity than any process change, since a designer reaching for a component that already exists isn’t making a tricky decision call. But a rushed system produces components people quietly refuse to use, and then you’ve paid for the thing and kept the silos anyway.

Finally, run design sprints ahead of development rather than inside it. In a dual-track setup the discovery work happens in parallel, so by the time engineers start building, the flows and the research behind them are already sitting there. With AI prototyping setups, you can work technical feasibility and validation of your solutions right into the discovery phase without tons of dedicated engineering resources.

And so engineers stop having to wait and designers stop trying to sprint after a train that left the platform without them.

Lead the conversation

Make the work visible, because a lot of the complaint is actually about opacity. A board people can see, a slack channel where sketches turn into wireframes and then into something clickable, a monthly showcase, the format matters less than the idea that people get to watch it happen. Visibility kills most complaints and misconceptions before anyone has to answer them.

When you do push back, bring receipts. Name the project where skipping research produced a feature nobody used, or the rushed flow that cost engineering two months of rework afterwards.

Make it safe to ship something imperfect on purpose. If your designers thinks every release is a referendum on their craft, they’ll polish the backend modal to the same standard as the booking flow, and you’ll have fostered the exact slowness you’re trying not to get blamed for.

The one thing that won’t work for you, is more explaining. Every meeting you spend defending your creative process keeps everyone’s idea of slow design exactly as wrong as it was when you walked in.

Hi, I’m Oscar - Founding designer at momondo, I’ve won a Material Design Award for Innovation, and I help design leaders succeed.

Subscribe now

Your CEO didn't decide design was overhead. You did.

Whenever an exec hears “design thinking”, they don’t picture the massive ROI that IBM got out of implementing it, or the weekend Airbnb’s founders spent saving the company with a rented camera on a hunch. Instead, they picture walls of sticky notes, and teams “ideating” for days without shipping anything. And it’s not a picture that fell out of the sky, but one we painted, one empathy-map workshop at a time.

We in the design community just love describing our work in the language of feelings – empathy, delight, human-centeredness. They’re all true, and every single one of them land in a CEO’s ear as a spend with absolutely zero line back to revenue. So what we really built was a vocabulary where we priced our own work as overhead, and then we go and act surprised when finance treats it that way.

That’s the surface problem, and it’s one we can actually fix. The deeper one sitting below that is that we happily got in that seat in the first place.

Your exec is rational – and that’s your whole problem.

I personally love the way Rory Sutherland always has a habit of pointing out things so incredibly obvious that nobody has bothered to say them: Creative people present ideas to rational people for validation, but the opposite never happens.

No CFO has ever walked into a design department holding a spreadsheet and asked for a creative look at it (and if he did, that’s a pretty big red flag). The traffic goes one way, and it has done so for as long as any of us have been working, and our grandparents too. Creative functions get the power of suggestion while finance, analytics and legal get the power of veto.

That asymmetry is the deciding factor for more product outcomes than any other methodology. Any idea that can’t be justified in advance dies in advance. This means the only survivors are the ones that walk and talk like the things already proven to work.

One of Sutherland’s go-to notions is that if an idea isn’t a bit absurd at first, then there’s no hope for it, because anything immediately sensible is already being done by your competitors. He’s borrowing the phrasing from Einstein, but he’s spot on about the business consequence. Sensible ideas are commodities, and they pass review precisely because everyone recognizes them, which is the same reason everyone – and definitely your competitor – is already shipping them.

So we have to accept that the vocabulary problem and the veto problem aren’t the same problem after all. Better words might get your budget approved, but they still don’t get you the right to be wrong.

Evidence is what gets you in the room.

You can’t just walk in the room and ask for things just on faith alone. You start with the translation, which is how you buy standing, and standing is then what you can spend later.

McKinsey scored 300 public companies on design over five years and matched the scores to the financials. The top quartile grew revenue 32 percentage points faster than their industry rivals and returned 56 points more to shareholders. That gap held up across medical technology, consumer goods and retail banking, so it isn’t a quirk of just one sector.

IBM sold design thinking internally as operational efficiency, which is why executives signed off. This way it spoke to margins rather than mood. (Most consultants still confuse operational excellence with process excellence. They’re wrong, and overpaid.) A study of IBM’s practice found a massive ROI and millions of dollars in net value gained over just a couple of years.

Intuit ran the same play under a different name, where Scott Cook came up with Design for Delight (D4D) as a way to find revenue by watching customers, and what eventually convinced the company came from a two-day sprint testing a hunch about QuickBooks licensing, showing that customers wanted single-seat licenses rather than their current five-packs, and by changing the packaging, they added $10 million in sales in the first year.

None of these companies or people found a better methodology than yours – they just found better words for the same one. And numbers. And every one of those numbers is a key to a door.

Nobody would have approved Airbnb’s Hail Mary.

In 2009 Airbnb’s revenue was flat and the founders were living on credit cards. They rented a camera, flew to New York, knocked on hosts’ doors and replaced the grainy listing photos with professional shots. Revenue roughly doubled that week.

Now try running that through a rational review, and you’ll find it doesn’t scale at all. The sample is a few dozen listings in one city, you’ve got no control group, no projection, and no claim you could make in advance for why better photography would move bookings. Any competent senior stakeholder would kill it in the meeting, and they’d be doing the right thing according to the books.

Co-founder Joe Gebbia put it like this:

"We had this Silicon Valley mentality that you had to solve problems in a scalable way... the first time someone gave us permission to do things that don't scale... it changed the trajectory of the business."

Permission is the word to notice there. The rare thing was never the idea, but the agency to run it without proving it first.

This is the asymmetry of creative ↔ rational doing its thing, and it’s why just learning the lingo on its own leaves you stuck. A design function that can only fund what it can pre-justify will produce competent, defensible and thoroughly forgettable work forever. It also gets to keep its budget, which is how the cycle survives.

Ask for a budget you can’t justify.

It’s not 4D chess, but it’s got a little bit of the flavor: First, you earn the credibility in their currency, and then you spend it buying back some veto-free ground.

Start by trading the vocabulary: User research is market validation, a prototype is risk reduction, and a journey map is a conversion analysis. Use the words your execs already think in and most of the initial hesitance goes away before you’ve even made your case fully.

Chop it down to size, and run one pilot rather than a whole program. Take a single high-visibility feature and give it a week of interviews and rough prototypes, tracking what you learn, what it moves, and what it would have cost to learn the same thing after launch.

Lead with other people’s numbers when you present. Finding relevant examples from your industry, like the ones I started with, because specific outcomes from named companies persuade in a way that theoretical benefits can’t.

Get your exec in the room while you’re at it. Invite them to sit in on a user session, since nothing changes an executive’s mind faster than watching a customer fail to buy the thing they’re trying to sell.

And then float what you’re really asking for. Once you’ve gotten the agreement, request a small budget with no business case attached, on the explicit understanding that most of what it funds will fail. You can frame it as portfolio logic, which is a concept every executive already accepts everywhere except when your work is involved. Ten percent of your capacity, spent on things nobody can justify in advance, is a really cheap insurance against a competitor finding the absurd idea first.

That last one is the actual ask – the four before that just make it sound like it came from an adult.

Where good ideas live.

Logic is a starting point that most companies also mistake for the finish line. All the interesting work that you should actually be doing sits where three things overlap: an economic insight that holds up, a solution that technically works, and a psychological framing of how someone then “experiences the thing”.

Most orgs are only staffed for the first two, leaving the third with no natural owner, no dashboard for it, and nobody in the room to speak for it when the decisions get made. And that gap is what design needs to be claiming, and it’s a considerably better pitch than asking to be included earlier in the process. (You’ve asked that a hundred times, haven’t you?)

AI is changing the price of being wrong.

The obvious objection is whether AI makes all of this moot, but it actually does the exact opposite.

When building gets cheap the cost of a failed experiment collapses, and the entire rational case for demanding proof up front weakens along with it. Veto power was always justified by how expensive it was to be wrong, and that expense is falling fast.

The risk now is that designers meet this moment the way they met the last one, talking about craft and taste while the business talks about outcomes and impact.

So, if once again, you go about describing the work that way, you’ll be overhead again inside a year. Instead, describe it as the discipline that decides where all the cheap, fast building gets aimed, and you become the constraint that matters, because everything else has stopped being scarce.

The work was never the problem, it was the words we used, and the words are entirely yours to change.

So take one project this quarter, run it as market validation, and put the money it saves in front of your exec. Then use the credit to ask for something you can’t defend. If the answer comes back no, you’ve learned what your seat is actually worth, which is worth knowing before you spend another five years negotiating for a slightly bigger one.

Hi, I’m Oscar - Founding designer at momondo, I’ve won a Material Design Award for Innovation, and I help design leaders succeed.

Subscribe now

I get the panic. When the first cameras appeared, portrait painters watched their whole trade seem to evaporate overnight. The wealthy stopped commissioning painted portraits the moment photography could capture reality faster and cheaper. Doom and gloom, right?

Except that’s not what happened. Photography didn’t kill painting, it freed it. Painters stopped trying to recreate reality exactly and started exploring abstraction, impressionism, and forms nobody had tried before. The medium found what it was actually for once it stopped competing on the thing machines did better.

The conventional wisdom says AI spells doom for professional designers. It has this backwards. Designers are becoming more essential than ever, just not in the way most people think. AI made the easy stuff cheap, and in doing so it opened a wide gap between generating a design and knowing which one is worth building. Which means the real question is whether you’ll spot your new advantage before your competition does.

Handle the flood

The flood of generative AI tools is reshaping what it means to be a product designer. The old playbook, where designers follow the double diamond to the letter, crafting pixel-perfect UI and polishing visual details, no longer applies. AI has made design creation cheap and easy for almost anyone. Your product manager creates wireframes over morning coffee. Developers vibe code interfaces during lunch. Your CEO, yes, the one who prints emails, generates a credible-looking design from a text prompt.

To a seasoned designer this feels like watching robots take over the assembly line. Equal parts fascinating and terrifying. If everyone can design, what’s left for me? you wonder, staring at yet another AI-generated mockup that looks suspiciously competent. Yikes.

But the answer isn’t producing more visuals faster. Sure, I can birth a prototype in Lovable or Figma Make faster than you can say “design sprint,” but claiming the design process now takes minutes is like saying cooking is just heating food. (I’m looking at you, Yotam Ottolenghi, and your recipe timings.)

AI is a tireless idea machine, a creative tornado that never stops spinning, and it has the evaluative capacity of a goldfish. It can’t tell brilliance from garbage, innovation from imitation, or a solution from a pretty distraction. Worse, the feedback loops these models run on keep reinforcing the bad outputs. Take a look at your LinkedIn feed if you need proof.

Become a strategic curator

This is where your value shifts into something more strategic. You move from creator to curator. While AI spits out infinite variations of every interface pattern you can imagine, you’re the one separating the valuable concepts from the attractive waste, deciding which ideas deserve development budget, tying them to business objectives, and making sure they solve human problems instead of just looking good in your portfolio.

Your edge now rests on three things. Discernment, recognizing what will actually work in messy organizational and user contexts rather than what looks good in isolation. Taste, the kind that tells a strong solution from a merely competent one, built up over years of pattern recognition. And critical thinking, the habit of questioning the assumptions under a request before they harden into expensive mistakes.

Because generative AI has the design taste of an algorithm, which is to say none at all. The tools keep getting better at producing a competent screen, and they’ll keep improving. What they don’t do is tell you which screen is worth building, for whom, and why. That gap is about context and consequence, not pixels, and it isn’t the kind of thing a bigger model closes on its own.

Experienced designers have something AI can’t replicate: the hard-earned ability to see what actually works in human contexts, and to understand why it works.

AI-generated outputs are like cover bands. Technically proficient but rarely soul-stirring. They reflect the greatest hits of their training data, which lands you designs that are competent but about as groundbreaking as vanilla ice cream. Prompt a model for a design right now and you’ll get something clean, confident, and oddly generic, the visual equivalent of the average of everything it has ever seen.

Nurture your inner skeptic

AI’s biggest blind spot is that it’s a yes-man in algorithm form. Ask it to design something and it’ll execute your prompt with the enthusiasm of a golden retriever. No questions asked, assumptions unchallenged, biases amplified. And you don’t have to look far for where that leads. When xAI loosened Grok’s guardrails to make it less “woke,” the result wasn’t sharper judgment, it was a run of offensive outputs and regulators knocking on the door. A model pours its energy into whatever you point it at, useful or catastrophic, without ever asking whether it should.

So designers step into their most important job yet, as the designated dissenter in the room. You’re not just evaluating outputs, you’re interrogating the premises behind them. You’re the one asking the uncomfortable questions:

• Are we solving the right problem for the right users?

• Who actually benefits from this design choice?

• What are we assuming about how users behave?

• How does this line up with the wider business objectives?

• What unintended consequences might come out of it?

You become the user’s advocate in a world where AI will happily optimize for engagement while ignoring whether people feel respected, understood, or shut out. And that job matters more every month, as the volume and speed of AI output climb.

Work with stakeholders and their cracks at AI

Your CEO just walked into your office with an AI-generated prototype that actually... doesn’t suck. Welcome to today, where senior stakeholders have become citizen designers, armed with generative AI and dangerous levels of confidence.

Your first instinct might be to build fortress walls around your design process, muttering about “amateurs” and “staying in their lane.” Don’t. That path leads to irrelevance faster than you can say “design thinking workshop.” Embrace the chaos instead. When a stakeholder shows up with an AI-generated concept, treat it like a rough sketch rather than a finished solution, and use the moment to start a real conversation.

So instead of resisting, do four things. Validate the thinking that got them there, without handing over your judgment. Pull out the useful parts of what they made, the bits that line up with user needs and the business. Point out the gaps they couldn’t have seen, the user considerations and technical constraints hiding under the surface. Then show how professional design work turns a starting concept into something you can actually ship.

Rather than being the gatekeeper who says no, become the translator who says yes, and here’s how we make it work for humans. That prototype Steve from C-suite just handed you? It might actually be the most coherent brief you’ve had all year.

Build better systems or drown

One more shift, and it’s the one that decides whether everything above actually scales. The explosion of AI-generated artifacts means a solid, scalable design system stops being a nice-to-have and becomes survival infrastructure. Without strong, intentional foundations, generative tools will flood your product with inconsistent experiences, redundant components, and enough technical debt to make your engineering team weep. Ask any developer who’s tried to wrangle AI-generated code without guardrails. The pain is visceral, and it isn’t just a tech-giant problem anymore.

That turns designers into system architects and stewards, building the frameworks and guardrails that let everyone, designers, engineers, stakeholders, and yes, AI itself, work at scale without wrecking quality or coherence. A strong system is leverage as much as defense, and it stretches your influence across the whole organization.

The painters who survived photography became the painters we remember.

Just as photography freed painters from the impossible job of competing with mechanical reproduction, AI frees designers from competing on speed and volume. The painters we celebrate now, Monet, Picasso, Van Gogh, found their value once they stopped trying to be human cameras.

Your role hasn’t disappeared. It’s turned into something more strategic, more influential, more central to whether the business wins. The flood of AI content isn’t destroying design, it’s showing who actually understands what design does beyond surface looks.

So take one project this quarter. Run it as the judgment layer rather than the pixel layer. Put the result in front of your CEO and let it make your argument.

Or keep running empathy workshops and wondering why the budget keeps going to engineering. 🫪

This is the first of 3 articles on this topic - find the other two here:

Hi, I’m Oscar - Founding designer at momondo, I’ve won a Material Design Award for Innovation, and I help design leaders succeed.

Subscribe now