How AI changed my work as a designer

Superpowers for a designer

AI changed my work as a designer massively. For the first time, I didn't feel like I needed to work with my hands tied behind my back. It truly felt like I had gained superpowers.

My client colleague and technical person recently said that it has been a huge paradigm shift to get me into coding. And because of that, it has helped shorten the gap between me and the product. That has been a total game changer, for me personally as a designer, but also for the whole team's ways of working.

Design, test and implement in the real product

Now I can do most things myself in the real product. I can design, I can test, I can iterate and implement right in the real product. Developers don't need to build everything from scratch. And that makes everything so much easier, because when I can test in the real product, there's much less for QA to find. I usually find the issues before they do.

Now in my workflow, I can design, iterate, test and implement in the real product, so the developers can focus on reviewing and building the back end. It means there are far fewer handoffs between me and the developer, and between the developer and QA. So it's much smoother and faster.

Where design actually happens

For me, it's much more about where design actually happens. There's a ton of debate about whether design should be done in Figma or with AI, with Claude Code or something. But I think the focus should still be on the solutions, the outcome and solving the right problems. If you focus on that, the answer comes naturally. You don't even need to think about which tool you should use. And for me, it has been with the help of AI.

As a designer, solutions and ideas have always come to my head almost like dreaming of them. So I don't need Figma precisely. I just need a place or a way to get them out. And for me, doing it with AI has become a very natural workflow.

It's so much easier to test everything, including edge cases, in the real product, because it's the same thing the end users use.

AI gave me time back

AI doesn't design for me, and it didn't change everything, but it gave me time back. So I can focus more on the bigger things, on solutioning and iterating, instead of pushing pixels in Figma like I did in the past. The human part didn't go away. It just moved. Now, instead of moving pixels in Figma, I can focus on iterating and testing in the real product.

What it took to get started

Picking up coding tools wasn't total magic. I already had deep product knowledge, which is really important. But at the same time, I had human developer teammates helping me in the beginning. That was really important at the start, because otherwise I would have gotten stuck in the wrong places. Now I could just move forward. And getting access to the real product, not a sandbox. Getting my hands dirty right away was one key thing.

No, it's definitely not always a smooth ride. Sometimes you ask Claude to move just the smallest thing, and it does that, but at the same time it does a million other things you don't want it to do. That really highlights that it still needs human judgment.

Basically, I don't give AI superpowers to change and do everything. I take the superpowers from AI to help me.

No fixed way of working

Between designers and developers, I don't think there's a fixed way of working anymore, and that's the real shift. Sometimes I start designing and implementing, and developers refine it after me. But sometimes it's the other way around: a developer builds the prototype first, and afterwards I do the UI/UX pass. So there's no fixed way anymore. It changes and evolves as we go. What matters is that nowadays it's truly shared. It's not just us designers vibe coding and developers cleaning up after us. It's a two-way street, and we develop things together as we go.

I think seeing things also helps get the best out of more technical people. Now that I can make working prototypes much more easily, it's also much easier to start discussing them with developers or POs. And that helps bring out their solutions and perspectives too.

I think what made it work for our team was that we share the same context. We have the same product knowledge. We share the same PR review skills and design system skills. So it's not my personal setup. It's something we build on together. And I would really highlight that you need to be ready to evolve, personally but also as a team, because nowadays things change so fast.

The bottleneck moved to reviews

Now that we use AI much more as a team, the bottleneck has moved. It moved from designing and developing to reviewing. Reviews take much longer. That's why, as a team, we develop different PR review skills and keep building on them. Personally, I've also been thinking about which developer skills I should learn. For example, I'm learning to make smaller PRs, slice them down and build umbrella PRs, so it's not such a heavy load for the developers. From talking with devs, I've understood that it's not a designer vibe coding issue. It's a more universal issue that has come with the speed of AI.

A racehorse with blinders on

Honestly, I don't even try to keep up with all the new AI stuff flying around. Before I started using AI at work, everything felt really overwhelming. Slack felt like a place full of new skills and tools I was supposed to learn and pick up. So I decided to learn the basics and then work like a racehorse with blinders on, and not get distracted by all the shiny things. That has been a huge help for me. I keep my eyes open, and I pick something up when I feel the need for it, not because it's trending or someone else is using it.

For example, my colleague recently built a Claude skill for synthetic user testing. User testing is really important in product development, but sometimes it's really hard to get those answers. So I immediately started using it, and it has been a huge help. It doesn't replace human input, but it gives you a good sparring partner to move forward.