INDUSTRY
Who Do You Expect to Please With That?
There’s an old joke about a man who is not well-endowed. That’s the question he gets asked. And his answer — grudgingly, honestly — is: myself.
He knows exactly what he has. He knows exactly who it serves. And to his credit, he doesn’t construct a narrative around it. His constraint explains his choice. It doesn’t excuse the outcome — but it accounts for the decision.
Every time the industry ships something the customer didn’t ask for and can’t fully access, it gets asked the same question.
The industry doesn’t have his constraint. And I question whether it has his honesty.
The accommodation cost
Roger Martin has observed that customers want what technology does for them — not the technology itself. And critically: they don’t want to pay an accommodation cost to get it.
Lululemon introduced advanced fabrics that flattered and performed. The customer got the benefit without adjusting to the material. Four Seasons deploys the latest in hospitality — but anywhere in the world, it feels like home. The innovation is real. The disruption to the guest is not.
The ones who win will bring the latest advances and require the least accommodation from the human.
The latest technology will feel familiar.
This is the standard GenUI should be held to. Not “is this technically impressive?” but “who is absorbing the accommodation cost?” If it’s the customer — if they’re reorienting, relearning, adjusting — the technology hasn’t been finished yet. It has merely been shipped.
What GenUI can generate — and what it can’t
User experience is not the interaction alone. It is motivation — interaction — outcome. The interaction is only the middle.
GenUI can generate the middle. It can produce an interaction that functions — that responds, adapts, presents options. That’s genuinely capable. But an interaction that functions is not the same as an experience that serves.
What lives on either side of the interaction is the human: what they bring to it, and what they need to leave with.
High-frequency interactions must become automatic or they drain attention at every use. High-criticality decisions need to be survivable if something goes wrong. High-compulsion moments — a ringing phone, an incoming notification — arrive when attention is already divided and the system is competing with itself.
These aren’t just categories of use. They describe what the person brings to the interaction and what they need to leave with. If the GenUI prompt doesn’t carry that context — if motivation and outcome aren’t designed in before execution is handed over — the interaction gets generated and the experience misses its potential.
Don Norman identified two gulfs that determine whether interaction succeeds. The Gulf of Execution — the gap between what you intend and what you ask the system to do. The Gulf of Evaluation — the gap between what the system produces and your ability to judge whether it’s right.
GenUI doesn’t close either gulf. It accelerates across them.
Throw it over without enough specification and the Gulf of Execution opens before the first line is generated. Receive the output without a clear standard to evaluate it against and the Gulf of Evaluation swallows whatever came back. Both gulfs have the same root: the work of defining motivation and outcome wasn’t done before execution was handed over.
GenUI didn’t create that gap. But it will find one if the prior discipline isn’t there.
The gap appears where it always appears: when the interaction asks more of the person than it should.
Confidence is the capacity that remains when interaction does not demand more than it should.
Confidence depends on four conditions: that the system behaves dependably, responds when you act, feels familiar enough not to require interpretation, and is safe enough to attempt when you’re uncertain. Familiarity isn’t the enemy of innovation. It’s the evidence that the innovation is complete — that the accommodation cost has been absorbed by the system, not transferred to the person.
When those conditions hold, people act. When they don’t, people compensate — they hesitate, check twice, choose the safer option, find workarounds. The interaction survives. The experience fails its potential.
The wrong constraint
There’s a principle in manufacturing worth borrowing: to optimize throughput, subordinate the system to the actual constraint. Not the most exciting one to solve. The real one.
Increasingly, the industry builds for what suppliers are pushing, what competitors are shipping, and what organizations want to be known for — and constructs a customer rationale around it after the fact. The room isn’t being read. It’s being assumed.
The customer is already responding. And the signal is being misread.
The push to return physical controls to vehicles is being interpreted as a preference for materiality. It isn’t. What customers are signaling is a need for predictability and legibility — controls that behave the same way every time, that don’t require interpretation at the moment of use. Physical is not the point. Confidence is.
Doug DeMuro recently reviewed the Lincoln Navigator and noted the steering wheel buttons — physical, present, tactile — but contextual. Ambiguous. Their meaning shifts depending on what the system is doing. You have to check before you act.
His verdict: you’ll get used to it.
Perhaps the most overused line in interface design. And the clearest sign that the accommodation cost has been noticed, named, and handed to the customer anyway.
When? is the first question. Why should they have to? is the better one.
The steering wheel sets the standard for every control attached to it: something you don’t have to look at or think about to use. That standard was earned over a century of refinement. It is the highest-confidence zone in the vehicle — not because attention is focused there, but because the interaction asks nothing of attention at all.
Every control added to the wheel has to earn that same standard. Not approximate it. Meet it.
A contextual button on the steering wheel doesn’t just fail that standard. It borrows against the confidence the wheel itself has already established. It seeds a question — do I know what this does right now? — in the one place in the vehicle that had no business introducing doubt.
The form is physical. The failure is the same.
The constraint in interaction design is not what can be built.
It’s what the person can access with confidence — in the moment they need it, without being asked to manage the system when they most need to manage something else.
The man in the joke had a constraint that explained his choice, and the honesty to own it.
The capability to do better is here. The question is whether we’re willing to read the room — or whether we’ll keep building for ourselves, constructing a customer rationale around it, and calling it service.