The AI Insider: Figma's native AI agent launches as revenue surges 46%

Figma combined two pieces of news on May 20–21: its native AI agent entered limited beta inside the collaborative canvas, and the company disclosed first-quarter 2026 revenue of $333.4 million, up 46 percent year-on-year.

The agent accepts natural-language prompts to generate designs, edit existing files, and iterate on layouts. Multiple agents can run simultaneously on the same canvas, letting human teammates and AI operate in the same multiplayer workspace. The underlying models are fine-tuned specifically for design contexts — they understand component structures, layout logic, and visual hierarchy rather than treating a Figma file as generic data.

The revenue figure provides context for the product launch. Figma is building its AI stack against growing pressure from Canva, Adobe, and a generation of AI-native design tools that did not exist before 2024. The company has assembled its AI capabilities in three stages: Claude Code integration via Anthropic, Codex integration via OpenAI, and now a first-party agent it controls directly. Owning the agent layer matters because it is the part of the stack that mediates how designers interact with all other AI capabilities.

For design practitioners, the competitive pressure that is driving Figma’s product acceleration also means that the tool they depend on is changing at a faster pace than at any earlier point in the company’s history. Features entering beta this quarter may become defaults within months, and adjacent tools are undergoing the same cycle simultaneously. Teams that evaluate AI features deliberately — rather than adopting everything as it ships — are better positioned to build coherent workflows rather than accumulating disconnected shortcuts.