AI is making it easier than ever to design the wrong thing beautifully.
Working across UX, product design and development, I see how quickly AI is changing the way we build digital products. We can now generate interfaces, create working prototypes, explore multiple solutions, and even build complete applications in a fraction of the time it used to take. And I think that’s a great thing.
1. Faster execution doesn’t mean better products
Here’s what I believe: faster execution doesn’t automatically mean better products. As AI takes over more of the execution, the value of designers is shifting from creating screens to making better product decisions.
When a polished screen costs almost nothing, polish stops being a signal. A beautiful prototype of the wrong idea is still the wrong idea. It just takes longer to notice, because it looks finished.
2. What the work is becoming
It’s becoming less about how efficiently we can design something and more about:
- Understanding which problems are actually worth solving.
- Challenging requirements instead of simply executing them.
- Thinking beyond individual screens and understanding the entire product experience.
- Balancing user needs, business goals and technical constraints.
- Knowing what to build, what not to build, and why.
None of these show up in a screenshot. But they decide whether the screens that follow are worth anything.
3. Generating options isn’t making decisions
AI can already help with many of these things, and it will only get better. It can summarise research, map a flow, list edge cases and argue both sides of a trade-off.
But generating options isn’t the same as making good decisions. Ten directions in an afternoon is a gift, and also a new problem: someone still has to choose, and own the choice. That means holding the user need, the business goal and the technical constraint in your head at the same time, and finding the place where all three are true. That overlap is where better products come from.
4. Knowing what not to build
The most valuable decisions I’ve made on my own products were mostly about saying no.
Cohivo started as a broad brief: an AI controller watching scope, budget, dates, capacity and dependencies across a whole portfolio. A strong vision, and a hard first purchase. After researching more than twenty scope-creep tools, I cut it down to one sharp job: stop scope creep before it eats your margin. AI could have generated every screen of the bigger version just as fast. It couldn’t have told me not to.
On OpenSty, the same thinking led to booking requests instead of instant booking. Instant booking looks like the more capable feature. But small hosts already work by confirming each guest themselves, and handing that control to software was the thing they feared most. The simpler choice was the right one.
The biggest opportunity with AI isn’t just doing the same work faster. It’s spending more time on the work that actually matters.
AI generates. UX decides.
I don’t think the future belongs to designers who can create the most screens. It belongs to those who can turn ideas into meaningful outcomes.
Use AI for everything it’s good at: the first draft, the tenth variation, the prototype you can put in front of a user by the afternoon. Then spend the time it saved on the questions it can’t answer for you. What problem is this solving? For whom? And is it worth building at all?