Figma: AI agent rolls out directly inside the collaborative canvas

Starting May 20, 2026, Figma is rolling out a native AI agent to eligible users in limited beta. The agent lives directly in the Figma Design canvas and responds to natural language prompts: designers can ask it to create new layouts, modify existing ones, generate design variations, or automate repetitive tasks. Unlike earlier AI features in Figma that operated through menus or separate panels, this agent interacts with the canvas itself.

The models underlying the agent have been fine-tuned for design contexts. This means the agent understands design elements, layer structures, and common workflow patterns rather than treating design files as generic data. Figma says teams can run multiple agents simultaneously on the same multiplayer canvas, assigning them to different tasks running in parallel.

The business context is notable. Figma reported Q1 2026 revenue of $333.4 million, up 46% year-on-year, and raised its full-year forecast to $1.42–1.43 billion. This growth is happening alongside intensifying competition from Canva, Adobe, and newer AI-native platforms. The agent launch deepens the value of Figma’s core product for teams already paying for it.

Figma’s chief design officer framed the announcement around direction rather than execution: “As building software gets easier, what matters most is setting direction.” The implication is that the designer’s role shifts from producing individual frames to orchestrating what the agent produces.

Expansion to other Figma products beyond Design is planned, though no specific timeline was given.