Three weeks ago I had a conversation with a smart, experienced designer who told me AI changes everything about UX. Her exact words. I said: my process hasn't changed.

She pushed back. What about Figma Make? What about generative UI? What about the fact that a non-designer can now spin up a working prototype in ten minutes?

Sure. But I've been starting in full-fidelity since before Figma existed. I've been skipping unnecessary process overhead since I was the loudest voice arguing against it on large UX teams. I've been going straight to vision — ignoring tech stack, ignoring constraints, ignoring the seven-step research framework — and using the work to start the conversation. That's how I've always worked. Nothing about AI made it revolutionary. It was just always the right approach.

She thought I was being defensive. I wasn't. I was being accurate.

"Everything Is Dead" Season

If you spend any time on LinkedIn, you know we're deep in another round of [X] is dead because of AI. Designers are dead. UX is dead. Figma is dead. The junior role is dead. (That last one might actually be true — but not for the reasons being cited.)

I've heard some version of this my whole career. When every CEO and PM suddenly "knew" how to design because they could click through a prototype. When design systems were going to make designers unnecessary because engineers could just pull from the library. When no-code tools were going to replace the whole discipline. The people predicting the death of design have always underestimated the same thing: the part you can't template.

The data is starting to catch up with the hype. A recent study found a 19% productivity decrease among experienced developers using AI tools — because they spend the time they saved fixing the output. Research is showing 55% of employers who laid off experienced workers to replace them with AI now regret it. The tribal knowledge walked out the door. The complexity stayed.

My Confession About Figma

I never liked it. I got proficient at it because I had to. But my preferred approach was to misuse it — go straight to the idea, skip the named layers, ignore the auto-layout ritual, don't get precious about the tokens. Get something in front of people and think out loud together.

The industry spent years treating that ceremony as the craft. Naming conventions, component hygiene, token architecture. I watched designers spend hours organizing files that would be completely reworked in two weeks. Large UX teams did not love me for going around it. I went around it anyway.

And then AI came along and automated all of it. Figma Make generates the layers, handles the tokens, builds the components. All that manual work designers were told was the job — an AI does it now, and faster.

But here's what AI cannot do: decide what to build. Figure out which problem is actually worth solving. Know when a feature harms clarity instead of adding it. Sit in a room with a team arguing over requirements and reframe the whole conversation with a sketch. Look at a workflow that makes perfect sense on paper and recognize it will never work for a real person.

That was always the job. The ceremony was the distraction.

The Street Musician Test

I've played in bands my whole life. Not as a trained musician — I never learned theory, never practiced scales with a metronome. I wanted to play loud rock and roll in bars and clubs with other people. You either sink or swim.

Last year I did an instrumental surf gig with a classically trained keyboardist. Brilliant in a practice room. On the gig, he was a mess — couldn't feel the room, couldn't recover when things went sideways, had never heard of the Rolling Stones. All the training. None of the judgment you only get from years of being on a real stage.

Design is the same. The designer who has spent years in messy, real, uncooperative products — who has argued with engineers, rebuilt flows the night before a deadline, watched actual users get hopelessly lost in something they designed — that person has something no amount of tool certification produces. Figma's own 2026 State of the Designer report says senior designers with strategic ownership are seeing increased demand even as AI takes over execution work. That's not a surprise. The execution was never the scarce thing.

The Long View

I've been in this field since before it had a name — human factors engineering, they called it, back when Photoshop 3.0 was a cutting-edge design tool. I've designed through the dot-com boom and bust, mobile-first, design systems, and now this. Every wave has the same shape: a new capability arrives, people declare everything before it dead, technology blindness sets in, and then the field slowly corrects back toward what actually works.

What actually works is a designer who knows what matters. Who can look at a product and see what's harming the user, not just what's technically possible. Who understands that simplicity is harder than complexity and restraint is a skill, not a lack of ambition.

AI hasn't changed that. It's just made the designers who were only ever doing the ceremony a lot more visible.

Paul Shellooe is a Principal Product Designer and Enterprise UX Architect with 25+ years of experience embedded with Fortune 500 engineering teams. Based in Lisbon, Portugal.