People are still looking for the right name for it. Some call it design engineering, some say AI-native design. I think of it as agentic product design: I still design the product, and AI agents build it with me while I direct, review and decide.
A year ago I would have called myself a product designer who prototypes. Now most of my days are spent in Claude Code, with Figma open next to it rather than in front of it.
The work itself starts the same way, with the problem and the people who have it. What changed is everything after that. I write a short brief, the agent builds a first version, and I react to something real in the browser or on my phone.
That’s where I’m faster. Ideas that used to sit in a backlog, or never got built at all, are working by the end of the day. That’s how I’m building Morning this summer, and Vibe alongside it.
It’s also where my work got better. Things that usually get cut at the end of a project, like dark mode, a sound, the way a screen moves, or what happens on a slow connection, now get done, because I can try them, see them and fix them myself. I can check every page on a real phone, in light and dark, before anyone else sees it.
And it has made me more creative. When trying an idea costs an afternoon instead of a quarter, I try more of them. Vibe started as one of those ideas. Now it’s something I can open on my phone and hand to a friend.
It isn’t magic. The agent will tell me, very confidently, that something works when it doesn’t. I’ve learned to ask for proof: a screenshot, a test, the page loading on the live site. Deciding what to build, what to cut and when it’s good enough is still my job.
My job feels closer to what I always wanted it to be: less time describing a product to someone else, more time making it and making it right.
Next, I want to write more about how I work this way, the briefs, the reviews, the mistakes, so other designers can try it too.