No more hiding behind the execution
We have always been able to hide. Behind the design. Behind the flow. Behind the pixels. Behind the output. This was what we were being judged on. If it looked good and worked well nobody asked how you got there. This is how the output functioned as a shield. For designers this peaked during the era of dribbblisation. The output was becoming a goal in and of itself, instead of a way to accomplish a goal. With AI tooling this is no longer what sets you apart. As we’ve seen in other posts, the output was never the hard part (Luke Wroblewski and Linear have both made this case). Some people make gorgeous looking shields, but aren’t functional, or made shields while no one wanted a shield, or made shields just for themselves, because they liked how they looked.
People have been hiding behind their output shields because it’s much harder to decide what kind of shield you’re making, why you’re making it and if it even has to be a shield.
What wasn’t shielded
As we’ve concluded that it was never about the output, we need to think about what wasn’t shielded. This is reasoning, problem solving and judgement, as I’ve talked about before (link to judgement post).
This overview, these designs, are the deliverables we’re looking for more and more these days. It’s the question: “do you actually know what and why you’re building?”. If you can’t explain this you’re not a designer, you’re a tool for a designer. You’re a worker. And that’s the layer we’ve automated and commoditised. If you feel like AI is going to take your job then you’ve never been a designer.
What AI requires from designers
Without the shielding of our output, it makes it even more clear that our first move shouldn’t be to open a tool but to start asking questions. I can’t believe I’m writing a blogpost on something I’ve learned 15 years ago in school. Now powered by new ideas the questions we should be asking are: what are the objects in this system, what can and want people to do with them, what does the business need to happen? Not as a methodology towards output, but as the output. Research is becoming more and more important and “the design before the design” will become our deliverable. The answers don’t just produce a document but a direction for prompting.
The test is simple: can you explain what you’re building to someone from engineering, a client, a marketeer or an other stakeholder, without showing a screen? Not the interface. The system. What it is, what it does, why it works the way it does. If you can’t, you’re not ready to open Figma. You’re not ready to pick up a pen. You need to think first.
Closing thought
The market is moving in this direction, but slowly. Figma surveyed 323 hiring managers in 2025 and found that 56% are increasing hiring of senior designers, compared to just 25% increasing junior hiring. Systems thinking ranked as a top skill requirement at 47%, almost as high as visual design at 58%. Their own conclusion: “When anyone can use AI to prompt their way to a prototype, organisations understand that design is what differentiates good products from great ones. As building products becomes faster, the ability to make the right decisions, shape a clear point of view, and create meaningful experiences becomes even more valuable.”
The same report also shows something else: designers who define craft as visual polish are happier in their jobs and report better business outcomes than those who define it as problem solving. The execution-focused designer wasn’t suffering when this survey was being done. Since then new tools and improvements in other tools might have changed this number. (And lets not forget this is a Figma survey, there’s a bias.). The shield still works, for now. What the data doesn’t show is whether that holds in two years, or five.
The designer focussing on output was always the one most vulnerable for replacement. Three years ago this was replacement by a better designer, now it’s replacement by AI. AI didn’t cause this, it has always been there and made it easier to do.