Where does AI fit in product design?

Artificial intelligence is on everyone’s mind, even if the conclusions we come to are vastly different. Yesterday at Config 2023, the Figma team talked about how they’re thinking about AI.

Here are my notes from the talk:

Design is solving problems for people. It evolves along side technology. We’ve seen this throughout its history: analog → digital photography, single player → multiplayer design process, etc. We can’t simply ignore advances in technology; we should instead remain curious, vigilant, proactive, and ultimately, optimistic.

The goal of software is bridging the gap between intention and action. App-based computing has been moving more and more toward task-based computing, simplifying the interface for complex interactions.

Where does AI fit into that?

The “atomic design” model focuses on small, rigid design elements that can be combined into increasingly more complex patterns to create full user interfaces and workflows. AI is really good at recognizing patterns and can help us move up that stack, quickly generating new and useful ways of combining atoms. This ideation stage is often slow, repetitive, and full of experimentation. This is where AI shines.

This approach raises the ceiling and lowers the floor so more people can participate in the design experience. Stakeholders can become design partners, and the design practice can benefit from collaboration, shared goals, lower egos, and great ideas coming from anywhere.

What does this look like, practically?

As a product, Figma is a community-centered design tool. As a company, Figma’s goal is building practical, useful things that help people work. AI is a platform, not a product, so the company has been thinking about where the AI platform is best integrated into the Figma product. To do that, they have to think about where best it fits in the design process.

Here’s the process:

Brainstorm → Design → Build

In the brainstorming phase, AI can help by generating, clustering, and summarizing. Creating icebreakers for team ideas. Making sense of large sets of information. Summarizing big sets of feedback collected on a FigJam board.

In the design phase, AI can help by creating variations, completing user flows, tapping into existing design system components to make intelligent component recommendations, and even chaining together multiple plugin actions to create even more powerful automated actions.

In the build phase, AI can help with handoff, generating code, matching design components to code components, and suggesting code changes based on design.

I don’t know what AI in Figma will actually look like, but everything I heard from the team led me to believe they see it as a tool to assist and support designers, not replace them.

Jesse Gardner

Up Next: Structures and incentives

← Full Archives

Design Systems Daily

Sign up to get daily bite-sized insights in your inbox about design systems, product design, and team-building:

New to design systems?
Start with my free 30-day email course →