Figma CEO Dylan Field: How AI Will Transform Design

Designers need to be founders. And it feels like we are in the MS-DOS era of AI—ten years from now we will laugh that all we had was a chat box.

Welcome, Dylan. Quick room check—lots of designers and a good number of founders. Let’s do the Figma story, AI and design, and founder advice. To start, where is Figma today, and then take us back to the beginning?

We are hybrid, around one thousand seven hundred people, and now have eight products after Config. It is surreal. We are shipping fast and hungry for feedback on things like Make, Sites, Draw, and Buzz.

Go back to nineteen-year-old you. How did this become Figma?

Evan and I were at Brown. We asked “why now?” and got conviction around WebGL over drones. We explored games versus tools, chose tools, and after many twists, committed around mid-2013 to what became Figma Design.

Did you see it as a company from the start or just a project with a friend?

Company from day one. Worst case, I learn working with Evan, go back to school. Teal Fellowship gave time without debt. Founders need to buy themselves time.

How did you stay motivated through the pivots instead of getting stuck in pivot hell?

Building with Evan was fun—we prototyped to think. We even tried a meme generator for a week; the thesis was right but the work felt wrong. Having a cofounder helped even out the highs and lows.

How did you get your first users, and when did you decide it was time to launch?

Cold emails, coffee with designers I admired, internships’ networks. Designers gave surgical feedback and we iterated. We did a VC-intro tour; conversion was tiny, but Notion and the company that became Coda adopted early. Launch sooner than we did and charge sooner. Today I push teams to slim epics and ship in one to three month bites.

When did you believe it could be huge?

Embarrassingly late. Microsoft told us Figma was spreading like wildfire and asked if they should block it since we were not charging. That was the wake-up call. Watch for product market pull—when users drag the product out of you.

How did you get comfortable seeking tough feedback?

I was a child actor—I auditioned and got rejected constantly. Rejection is data. You should want the data.

Design seems very front-and-center lately—Airbnb, Netflix, Apple’s liquid glass, your Config launches, and OpenAI buying Jony Ive’s firm. Why now?

Design’s importance has compounded for a decade, moving from lipstick-at-the-end to shaping how it works. As development gets easier with AI, the differentiator is craft, detail, and point of view. On OpenAI and Jony Ive: some moves you do not understand at first; assume there is something to learn before dismissing.

Tell us about the AI-focused launches at Config.

We watch emergent behavior in Figma Design and spin out dedicated tools so Design can stay focused. Draw deepens vector expressiveness. Buzz powers mass asset creation from brand templates. Sites lets you ship sites you design. Make goes from prompt to app—amazing for rapid prototyping and throwing ideas away faster.

Design and development are blurring. How is that changing your process?

It is about faster iteration and cheaper tests. Make accelerates zero-to-one. Models are currently stronger at early-stage prototyping than late-stage, complex codebases. AI is pushing generalist workflows across product, design, research, and engineering.

What comes after chat as the dominant UI?

This feels like the MS-DOS era of AI. The hard problem is exposing capabilities. Midjourney’s Discord and Meta’s public-by-default moments taught people what is possible. UIs will get more contextual across apps, and surfaces will multiply—think glasses and ambient displays—so consistency across contexts becomes the design challenge.

What is design’s role in AI research and model building?

Embed designers in research. Treat users and their problems as the anchor. Pair qualitative research with deep model work to surface what people are actually trying to do. It makes you faster and more right.

How does the designer’s role evolve over the next decade?

Designers get more leverage. More designer-founders and design GMs. Everyone will contribute to design, but designers will lead, curate, and set the systems and point of view.

Most interesting internal uses of AI at Figma?

Design and product should own evals. Researchers and engineers do not always have the closest line to the end user. Put designers at the center of evaluation.

If you were in your twenties today, what skills would you focus on?

Keep the fundamentals: learn broadly, think critically, relate to people. And do not substitute AI companionship for real relationships. AI boyfriends and girlfriends at scale would be a societal self-own.

What was the most fun period building Figma?

Right now. We have incredible people, tons of ideas, and the challenge is choosing the right ones.

I am a solo product engineer using Cursor for code and design, and looking at open source like Penpot. Is Cursor a competitor, and should Figma get more open and developer-friendly?

I do not see Cursor as a competitor—we launched our MCP server to get Figma designs into Cursor, Windsurf, VS Code, and others. The first AI generation is rarely the winner; the differentiator is the craft loop to make it good. On open source, we announced the acquisition of Payload CMS today. We want to support developers and open ecosystems more.

You said you productize observed behaviors. How do you watch repurposing and structure emerging use cases?

Mix of signals: support threads, qualitative sessions, watching people work, product data, and intuition—then test hypotheses. We started narrow on product design and expanded. On ethics, we are mindful, especially when using third-party models. On the horizon, more people will project consciousness onto AI, which creates hard HCI questions. We are at time—thank you all, and good luck on whatever path you pursue.