Figma launches tools to steer AI agents and standardise brand design rules

Figma has introduced new features designed to help organisations control how artificial intelligence agents interact with their design systems and brand guidelines. According to TechCrunch, the design platform announced these tools during its first Config event in India, aiming to bring more predictability to AI-generated outputs. These updates address the challenge of ensuring that automated agents adhere strictly to a company’s specific visual and structural rules.

The core of this announcement is Figma Specs, a Markdown-based editor that allows users to define their design language in a simple, structured way. This document serves as a central reference point for both human teams and AI agents working within the platform. By consolidating design libraries, code libraries, and guidelines into one location, Figma aims to provide agents with the necessary context to follow brand rules accurately while creating new assets.

What we know

Figma Specs can exist as a standalone file or co-exist within the design canvas itself. Beyond acting as a text editor, it allows users to reference other Figma elements, such as items on the canvas, FigJam diagrams, code layers, and existing designs. This centralised system enables agents to refer to a single source of truth, reducing the likelihood of non-deterministic AI outputs that deviate from established brand standards.

The company also expanded access to its agent on canvas feature, making it available to a wider user base. Given the inherent unpredictability of AI, Figma is providing these steering tools to guide the technology toward more consistent and predictable outcomes. This shift reflects a broader strategy where Figma positions itself as the place where software teams decide what to build and ship, leveraging AI’s capacity to generate code for discovery and design purposes.

Additionally, Figma has moved its code layers feature from private beta to public beta. Originally announced in June, this feature allows users to add coding directly to designs on the canvas for more interactive prototypes. Users can now ask the agent to turn a design into code or modify parts of the code to iterate on a prototype, bridging the gap between visual design and functional development.

Figma’s chief design officer, Loredana Crisan, noted to TechCrunch that there are 17 times more lines of code being produced today than in the past, yet only 1.3 times more releases. She explained that much of this code is built to discover what a product is, rather than to ship, highlighting the role of AI in the design and exploration phase.

The platform also announced support for over 20 new tools via its Model Context Protocol (MCP). According to TechCrunch, these integrations include Open Code, Dia, Grok Bot, and Indian AI startup Sarvam, further expanding the ecosystem of agents and tools that can interact with Figma’s design environment.

Defender's checklist

• Review how AI agents are currently accessing your design libraries and code repositories.
• Ensure brand guidelines are documented in a structured, machine-readable format like Markdown.
• Monitor AI-generated outputs for deviations from established visual and structural rules.
• Verify that third-party agents integrated via MCP comply with your organisation’s security policies.

The background in plain language

Organisations using AI for design face the challenge of maintaining brand consistency. AI agents can be unpredictable, often generating outputs that do not align with specific style guides. To address this, teams usually create a centralised source of truth for design rules. This document acts as a reference, ensuring that both humans and AI follow the same standards. By defining clear parameters, companies can steer AI towards predictable results, reducing the need for manual correction and ensuring that automated outputs match the intended brand identity.

Background: AI agents

AI agents are the software entities that perform specific tasks using instructions from large language models. Agentic AI is the architectural approach that grants these agents autonomy to plan, reason and act. You use agents as tools within agentic AI systems to achieve complex, multi-step outcomes without manual oversight at every stage.

Read the full guide: AI Agents vs Agentic AI: The Distinction That Matters

What to do and how to stay safe: Figma

• Audit your current design systems to ensure they are structured and accessible for AI interpretation.
• Implement a centralised repository for brand guidelines, including visual assets and code standards.
• Test AI-generated designs against your brand rules before publishing or sharing them externally.
• Once the vendor provides an update, review new MCP integrations for security and compliance risks.

General security guidance from the TechWire Beat newsroom. It is not confirmed advice from the organisations named in this story.

Frequently asked questions

What is Figma Specs and how does it help with AI?

Figma Specs is a Markdown-based editor that lets users define their design language, serving as a reference for teams and AI agents to adhere to brand rules.

Is the code layers feature available to all users?

The code layers feature is currently in public beta, allowing more users to test adding coding to designs on the canvas for interactive prototypes.

Which new tools are supported via Figma’s MCP?

Figma supports over 20 new tools via its MCP, including Open Code, Dia, Grok Bot, and Indian AI startup Sarvam.