Go back six or seven years. Figma came onto the scene and changed everything — and it deserved to.
Sketch had a real problem. It refused to evolve around what designers and product teams actually needed. No seamless way to prototype and gather feedback in the same place. Figma looked at that gap and solved all of it. Prototyping. Stakeholder sharing. Styling handoff. Version control. Real-time collaboration. All under one roof.
Sketch never recovered.
So here's the question: is Figma about to be on the wrong side of that same story?
Where Figma's value actually lived
Figma won because it compressed the entire design-to-feedback-to-handoff loop into one place. Your prototype lived where your stakeholder comments lived where your dev specs lived. That was the moat.
Now look at each one of those pillars and tell me they're not all being disrupted at once.
The Tony Stark era
I've started calling this the Tony Stark era.
I can sit down, have a conversation with Claude, think out loud — and use that conversation to build a markdown file that becomes the source of truth for an entire product. Design intent. User flows. Component logic. All of it, structured, coming out of a dialogue with a brilliant collaborator.
From there, I can build a full working prototype from a well-written prompt. What used to take days — sometimes weeks — now takes hours. Iterating on the design is seamless if you prompt well. You're not pushing pixels anymore. You're having a conversation with your product.
One of Figma's core value propositions, gone.
The handoff problem basically disappears
Handoff has always been where things go wrong. Specs misread. Interactions misunderstood. Time lost going back and forth.
Now I can use Claude to write requirement documents, map full end-to-end workflows — I've shared some of my favorite prompts for this on my AI Methods page — and hand generated code directly to a developer who can ship it.
The handoff isn't a handoff anymore. It's a head start.
So where does that leave Figma?
Figma isn't irrelevant tomorrow. But the writing is on the wall.
Figma's value was always compression — taking a messy multi-tool workflow and making it one place. LLMs like Claude, combined with VS Code and Claude Code, are compressing things even further. The entire pipeline from idea to prototype to documentation to delivery is collapsing into a conversation.
Sketch lost because Figma solved the problem better and changed where the work happened.
Figma might be facing the same thing. Not a better tool — a completely different paradigm.
The question isn't whether Figma needs more AI features. The question is whether a visual canvas is even the right starting point anymore.
I'm not sure it is.
Still designing. Just not always in Figma.