Figma has added generative AI features to FigJam, aimed at moving users from a blank page to a usable board in seconds. The company frames the work around practical problems rather than novelty: helping people visualize ideas and plans instantly, suggesting best practices, and automating tedious tasks so teams can focus on the bigger picture.

Using a prompt, users can generate templates for weekly team syncs, brainstorms, and reviews, or produce visual timelines and organizational charts for planning. The prompts are built on best practices and everyday use cases, and they are customizable so they can meet users at whatever point they are in their workflow.

Lowering the floor in practice

Vice President of Design Noah Levin described the goal as lowering the floor and raising the ceiling—making the product more accessible while expanding what's possible. A meeting request is the canonical example: say "Hey, I need a meeting with four people," and FigJam generates an initial template for that meeting. Because people already express intent in plain language through dialogue and texting, anyone can jump in and feel comfortable.

The features also target what the team calls the "blank canvas problem," where people freeze when faced with a new file. FigJam is positioned as an intuitive brainstorming partner that helps users reach next steps faster: summarizing a complex meeting or synthesizing ideas into buckets can take up to an hour manually, and AI can do that work instead.

Levin noted that while a visual space clarifies concepts and aligns teams, not everyone is fluent in design software. AI in FigJam is intended to lower that barrier for people across all disciplines, not just those with design backgrounds.

Defining a good board

Chief Product Officer Yuhki Yamashita framed the template work with a question: what if two of Figma's product and design managers, Jenny Wen and Emily Lin, were tasked with creating a FigJam board—what would it look like? That question guided the approach.

Product Designer Natasha Tenggoro said the team defined the elements of a good board—meaningful content arranged logically, effective use of widgets, and stickers adding a visual dimension—and then packaged those outcomes into core principles that could apply across a wide range of user needs. That packaging remains a critical part of working with AI, with the goal of making output as useful and intuitive as possible.

Engineering groundwork and a new writing style

The engineering effort built up gradually over the past year. Software Engineer Michael Bullington credits Engineering Manager Dan Mejia with championing the potential of AI to solve the blank-canvas problem, even before it was clear what the technology could do. Work on Jambot—which began as an idea at a recent AI hackathon and brought the generative power of ChatGPT to FigJam to help users ideate, summarize, and riff in a multiplayer canvas—paved the way for the current features.

To define what counts as "good," the team worked with design advocate Kaitie Chambers and product manager Albert Song, both experienced in creating templates and facilitating meetings in FigJam, to vet outputs and give feedback. UX writers helped establish a "FigJam writing style," addressing questions like what makes a concise paragraph or a good summary.

A discoverable, multiplayer approach

Tenggoro said the design had to be immediately understandable even to someone who has never used AI. For a visual tool, that meant reliable and discoverable entry points plus creation paths that lead users toward good outputs. The team settled on gradually building prompts through recommendations—if a user says they're setting up a meeting, FigJam might suggest an icebreaker, an agenda, and a brainstorming activity.

Levin framed the strategy as "making AI multiplayer," consistent with collaboration being central to Figma's products; the goal is accelerating what teams can build together and turning concepts into collective action. Tenggoro called the features a paradigm shift: a text-based tool familiar to non-visual communicators that lets them create something beautiful and useful in seconds.

Looking ahead, Bullington said he is excited about the growing machine learning platform team and about models that think visually, noting that many existing LLMs are largely text-oriented with a bias toward code—an area where Figma's strengths apply.