Why Designers Still Matter in the Age of AI

For months now, social media has been flooded with posts claiming that traditional UX/UI design is dead, that Claude Code (or whatever tool is trending this week) has replaced designers, and that anyone still using Figma is falling behind. Sure, we can't ignore the value AI brings in quickly generating ideas, speeding up the design process and handoff — but I still think designers are very much needed.

AI is a tool, just like Figma

About seven years ago, Figma showed up, and as a young designer who had been working primarily in Sketch, I resisted the idea pretty hard. I didn't like that it was browser-based, that my boss or client could jump into my file and watch me work, and I couldn't wrap my head around how Figma handled components. Looking back now, I know Figma is a much better tool — and maybe in seven years I'll say the same about some AI tool. But they're just tools. The thinking behind the design has stayed the same, regardless of how you execute it.

AI doesn't understand people

The idea that Claude has replaced designers falls apart the moment you realize AI doesn't understand fundamental human experience. It doesn't get what users actually need, what frustrates them, or the context in which they use a product. Good design requires thinking — and that's something tools haven't managed to replace yet. If we shift focus from user experience to the designer's actual workflow, there's still a huge chunk of the job that can't be handed off to AI. A big part of my career has been spent on corporate projects, where every new feature goes through a whole process — stakeholder meetings, business analysis, technical analysis, understanding the full context, and there are always external blockers along the way. All of that requires human judgment. The actual designing in Figma, Claude, or whatever tool — in that kind of environment, that's maybe 30% of the job. Good design was never (only) about translating briefs into screens.

So what is AI actually good for?

There's no denying that AI has changed the way we work. The biggest value I see right now is speeding things up — whether that's something as simple as replacing placeholder text with something realistic, or putting together a fully functional prototype in code for user testing. It made development handoff a lot easier in a lot of cases. It works as an idea generator too, but a prompt is only as good as your understanding of the problem you're trying to solve.

AI has democratized tools, but not experience. Anyone can generate screens that look decent — and sometimes that's genuinely enough. But the gap between something that looks decent and something that actually solves a real problem is significant. What rarely gets talked about is that AI doesn't learn the way we do. Every project I've worked on taught me something — how a certain type of user reacts to a certain pattern, what works in theory but falls apart in practice, when to cut a feature instead of adding one to make the experience better. That's intuition AI doesn't have and won't develop just from generating screens.

Tools change. Thinking doesn't.