Dylan Field on scaling Figma and the future of design

Executive overview

As AI makes software development faster and easier, design becomes the primary differentiator. Dylan Field, Figma's co-founder and CEO, traces how Figma was built through relentless user feedback and deliberate constraint, then argues that the same discipline applies to how founders should think about AI, interfaces, and the expanding role of designers.

Figma today has 1,700 people and eight products — each spun out of observed user behaviour in the core tool. The lesson: watch what people do, strip the roadmap down, and ship earlier than feels comfortable.

The companies that win in the AI era will be the ones that treat design as a core differentiator, not a finishing step.

From WebGL to Figma: the early years

• Figma began as a broad exploration of what could be built with WebGL, starting in earnest August 2012.
• Early pivots included a meme generator — right thesis, wrong product; both founders were ready to quit after a week.
• A co-founder is a forcing function: highs and lows cancel out and keep momentum alive.
• The Thiel Fellowship provided two years of runway — not just money, but time to avoid a premature kill decision.
• First users came from cold emails to respected designers and introductions through investors; conversion was extremely low but feedback was high-quality.
• The inflection point came when Microsoft flagged Figma was spreading internally and asked why Figma wasn't charging — five years in.

On launching and roadmap discipline

• Launch as early as possible; charge money early to validate whether people will actually pay.
• Figma launched late because feedback was clear the product wasn't ready — but in hindsight the team should have scaled faster to ship sooner.
• When a team presents a long roadmap, the first question should be: how do we make this more bite-sized and test it sooner?
• Constraints breed creativity; the internal cadence target is one to three months, never nine or twelve.
• Product-market pull is misread: obsessive feedback and feature requests are a signal to double down, not a sign the product isn't ready.

Design as differentiator in the AI era

• As development gets easier and faster, design — craft, attention to detail, point of view — becomes the primary competitive edge.
• Airbnb explicitly named design as their differentiator; the OpenAI acquisition of Jony Ive's company signals the same recognition at scale.
• AI is better suited to early-stage prototyping and zero-to-one work than to mature, established codebases.
• The current chat interface is the MS-DOS era of AI; the unsolved problem is how to expose model capabilities to users who don't know what's possible.
• Future interfaces will be more contextual, more ambient, and distributed across new surfaces — glasses, displays throughout daily life — requiring designers to think across a broad spectrum consistently.

Figma's product strategy

• Each new product (FigJam, Slides, Draw, Buzz, Sites, Make) was spun out of behaviour already happening inside Figma design.
• Keeping new use cases separate preserves the core tool's simplicity; one plus one should equal three, not 1.5.
• Figma Make (prompt to app) is already changing internal prototyping — fast iteration and cheap idea disposal.
• Signals for new products come from support requests, qualitative interviews, data science analysis, and social media — combined into intuition and testable hypotheses.
• The acquisition of Payload CMS signals a move toward open-source and developer-friendly workflows.

The evolving role of designers

• Designers will have far more leverage as AI lowers the cost of execution; design value will continue to rise.
• The number of designer founders will multiply — Brian Chesky and Kari (Linear) are early examples of a growing pattern.
• Most people in a company will contribute to design; designers will curate, lead, and set the system.
• Embedding designers in AI research teams is critical: they contribute intuition about how users think, which researchers trained in abstraction often lack.
• Designers should lead evals for AI models — they have more user contact than engineers or researchers and understand end-user intent better.
• The Steve Jobs framing applies: design is how something works, not how it looks; that makes designers central to model development.

Advice for founders

• Seek rejection early and often — it contains data, and the discomfort of hearing "no" is less costly than building the wrong thing.
• State your ICP narrowly at the start; Figma began with product designers at companies that cared about design, not the broad world.
• When pitching to an angel investor you respect, send a Loom first — async, low friction, and shows respect for their time.
• The startup cycle is always: identify what you're doing most of → find someone (or AI) to help → acquire the resources to do it.
• Keep the simple things simple and make the complex things possible — the most repeated design principle at Figma.
• Maintain real human relationships; do not substitute AI companionship for genuine social connection.