Figma’s AI agent strategy breaks from rivals

Figma launched an AI agent for product design on Wednesday, embedding the tool directly into its canvas and connecting it to user workflows, components, and brand standards. The agent can generate and edit design layers, automate busywork like component setup and bulk edits, and allow teams to work with multiple AI assistants in parallel without switching tools. The move positions Figma against rivals like Anthropic and OpenAI in an emerging battle over AI adoption in the creative design market.

Amid a plethora of AI-powered design tools, Figma is embedding agents where designers will actually use them. On Wednesday, the company unveiled an AI agent built specifically for product design, embedded into the Figma canvas and connected to user context. Rather than just embedding a simple chat interface, the company said this agentic canvas is a “natural extension” of the way design teams already work. In a press release, the company said the agent was built to understand Figma workflows, such as components, design system logic and brand standards. - Figma’s agent can help designers generate and edit design layers, generate new directions with “infinite canvas,” and get real-time feedback. The agent is a natural language tool, aimed to “ lower the barrier to entry for anyone who wants to design and build,” the company said in its release. - The agent calls on fine-tuned AI models trained to understand product design, patterns and Figma-specific use cases. Teams can also work with multiple agents in parallel, and navigate between AI-assisted design and manual manipulation without switching tools. - Because the agent is connected with the existing team context, product designers can automate busywork such as component setup, layout fixes, and bulk edits. “As building software gets easier, what matters most is setting direction: deciding what to work on, how it should function, what the experience should feel like,” said Loredana Crisan, Figma’s chief design officer, in a statement. The agent is now available in beta and will gradually roll out to Figma Design paid users in the coming weeks, eventually expanding to other Figma products. It’s the latest in Figma’s broader AI transition, which includes the natural language “prompt-to-prototype” tool Figma Make, advancements to its MCP server, and the ability to connect third-party coding agents. Our Deeper View Though design is part of Figma’s DNA, it’s not the only company that’s paying attention: Anthropic launched Claude Design in mid-April https://www.anthropic.com/news/claude-design-anthropic-labs , allowing users to collaborate with Claude to create “polished visual work.” Meanwhile, both OpenAI https://chatgpt.com/apps/canva/ and Google https://ca.finance.yahoo.com/news/canva-brings-brand-editable-design-174500985.html?guccounter=1&guce referrer=aHR0cHM6Ly93d3cuZ29vZ2xlLmNvbS8&guce referrer sig=AQAAADoMMkxX-hfJl0b99EXKj4GE1VwTVOtUqLk pj0mg8zK87Bo4n1kNlVT9I6Qj5Bl8UrlqyQGs mPpHtY-QD66KX2 Ihaf70BX1vBWFfEms0ZjkNyRx3kFGCF5jbn3xvADL2c s8EBIb LX7ayEb8QeYIjJUXNUNPHepiHw2Hca5x collaborate with the popular design tool Canva. As AI firms have long fought over the hearts and minds of software developers with AI coding tools, design may become the next battlefield for AI expansion in business. However, given that design is inherently a creative field, deploying AI in this market could raise the same ethical qualms that arise when AI is used in fields like filmmaking and music. While coding was a natural starting place for AI adoption, it remains to be seen whether design will have the same uptake.