Figma is teaching AI agents your brand rules: Specs and design context

On 14 October 2026, at its first Config event in India, Figma announced a set of features defining how AI agents work on its platform. The common thread is clear: agents now produce design, but without knowing a brand's rules their output is unpredictable. Figma is trying to close that gap.

Specs: writing down the design language

The new Figma Specs lets teams write their design language in a Markdown-based editor. A Spec can be a standalone file or live on the canvas. It can link to canvas elements and embed FigJam diagrams, code layers and designs. Crucially, both team members and agents use it as a reference while designing.

Design context management

Figma also introduced a system that centralises design libraries, code libraries and guidelines. Agents draw on it to follow a brand's rules. In Figma's words, the goal is to make AI-generated work more predictable.

Code layers move to open beta

Code layers, announced in private beta in June, are now in open beta. Users can ask the agent to turn a design into code or edit part of the code to iterate on a prototype. Figma's MCP support also expanded with more than 20 new tools.

Figma's chief design officer Loredana Crisan told TechCrunch that 17 times more lines of code are produced today than in the past, while releases have grown only 1.3 times. Much of the new code is written to explore product ideas.

What it means for brands

The message is: a brand guideline is no longer a PDF, it is a data source. Brands with a written, clear, current design system will get consistent results from AI tools. Brands with vague rules will meet a different identity in every agent output. The value of identity work is not shrinking; it is growing.

If you want to talk about how your brand should adapt to these shifts, request a quote.

Source: TechCrunch, 14 October 2026