# How to choose a design system for AI agents Selection and fit. Reviewed November 9, 2026. https://bestaireadydesignsystems.com/how-to-choose-a-design-system-for-ai-agents/ ## Answer in brief Start from what your agents will build and in which stack, then shortlist the public systems that fit it. Rank the shortlist by readiness: design tokens as data, an official server, docs as text and guidance for agents. Then read the conditions behind each Yes, such as a sign-in, a tool not yet generally available or a deprecated library. If none of them can carry your brand, the right design system for AI agents is your own, made AI-ready. Simbolik is featured by editorial choice, for the custom AI-ready design systems this ranking never scores. Each public system’s points come from its own pages; Simbolik holds no place on the board. ## Start from what your agents will build A design system is chosen for a product, and an agent only inherits that choice. So before you look at any system, write down which agents your team runs, what they make (admin screens, marketing pages, a mobile app) and in which stack. Most public systems are tied to one. Ant Design is React for data-heavy tools. Shopify Polaris is for apps inside Shopify. Salesforce Lightning is for Lightning Web Components. The GOV.UK Design System is for UK government services. A system outside your stack is not a candidate, whatever its score. ## Rank the shortlist of AI design systems by readiness With the shortlist made, read each system’s 4 ratings on the [best AI-ready design systems, ranked](https://bestaireadydesignsystems.com/) and ask what each gap would cost your team. - **No design tokens as data.** The agent copies values from CSS or guesses them, and you export the design tokens yourself. - **No official server.** The agent works from files alone. That can be enough when the docs are good: Microsoft Fluent 2 has no public server and still scores 5. - **No docs as text.** The agent reads web pages built for people, or falls back on what it remembers from training. - **No guidance for agents.** You write the rules yourself, before the first prompt. A gap you are ready to fill is cheaper than a system that does not fit. ## Read the conditions behind each Yes A Yes says the part exists on the makers’ own pages. It does not say it comes free of conditions, and some of those decide a choice: - **Carbon’s** hosted MCP server asks for an IBMid sign-in. - **Atlassian’s** skills for agents are marked private to Atlassian; its CLI, server and full-text docs are public. - **Salesforce Lightning’s** agent tools sat inside Salesforce’s general MCP server until Salesforce retired it on November 2, 2026, so an agent now reaches SLDS through skills alone. - **Chakra UI’s** server has tools for its paid templates that need a key. - **Shopify Polaris’s** React package is deprecated in favor of its web components, and its server covers all of Shopify, not Polaris alone. - **Material 3 for the web** is in maintenance mode, according to its own repository. Open the pages behind your shortlist’s ratings and look for the same kind of line. ## Adopt, adapt or build Three routes come out of the shortlist. **Adopt** a public system that scores 7 or 8 and fits your stack, and accept its look; your agents get the most help for the least work. **Adapt** one by theming it; the agent keeps the maker’s docs and server, but they describe the maker’s defaults, so every change you make needs guidance of its own. **Build** your own when the output must look like your company and no one else; it can be AI-ready from the first file. The [cost and scope of making a design system AI-ready](https://bestaireadydesignsystems.com/cost-of-making-a-design-system-ai-ready/) sizes that route, and [Work with Simbolik](https://simbolik.ai/service-ai-systems.html) is one way to have it built. ## Test the choice before you commit A score is read from documentation. Your agents will meet the real thing. Put the front runner through [a bounded pilot with an AI agent](https://bestaireadydesignsystems.com/ai-agent-pilot-with-a-design-system/) on 2 or 3 real screens before anyone migrates a page. ## Questions & answers ### Which design system works best with Claude Code or Cursor? None is tied to one agent. The official servers on the board use MCP, which many agents speak. shadcn/ui, for one, documents its server for Claude Code, Cursor, VS Code and Codex. Choose on readiness and fit, then check that your agent is on the server’s list. ### Should we switch design systems to get a higher score? Rarely. Moving every screen to a new system costs more than closing the gaps in the one you have. Score your own system first and fix its cheapest gaps; switch only when the stack or the look has to change anyway. ### Does theming a public system keep it AI-ready? Partly. The agent still reads the maker’s docs and server, which describe the maker’s defaults, not your changes. Every choice you make on top needs its own guidance, or the agent will drift back to the defaults. ## Sources for this entry 1. [Carbon MCP, onboarding and setup](https://www.carbondesignsystem.com/getting-started/carbon-mcp/onboarding-and-setup): Carbon’s hosted MCP server, which asks for an IBMid sign-in. Checked November 8, 2026. 2. [Atlassian Design System, get started with AI](https://atlassian.design/get-started/with-ai): Atlassian’s MCP server and CLI, with its agent skills marked private to Atlassian. Checked November 8, 2026. 3. [The Salesforce DX MCP server’s end-of-life notice](https://github.com/forcedotcom/mcp/issues/46): The retirement of Salesforce’s general MCP server, and the SLDS tools in it, on November 2, 2026. Checked November 11, 2026. 4. [Chakra UI, MCP server](https://chakra-ui.com/docs/get-started/ai/mcp-server): Chakra’s official MCP server, whose tools for its paid templates need a key. Checked November 8, 2026. 5. [@shopify/polaris on the npm registry](https://registry.npmjs.org/@shopify/polaris/latest): The Polaris React package marked deprecated, with Polaris web components named as the way forward. Checked November 9, 2026. 6. [Material Web on GitHub](https://github.com/material-components/material-web): Google’s Material 3 web components, which the repository says are in maintenance mode. Checked November 9, 2026. 7. [shadcn/ui, MCP server](https://ui.shadcn.com/docs/mcp): An official server documented for Claude Code, Cursor, VS Code and Codex. Checked November 8, 2026. ## Continue this reading path - [A procurement brief for an AI-ready design system](https://bestaireadydesignsystems.com/ai-ready-design-system-procurement-brief/) - [A bounded pilot with an AI agent and your design system](https://bestaireadydesignsystems.com/ai-agent-pilot-with-a-design-system/) - [What makes the best AI-ready design systems, the category guide](https://bestaireadydesignsystems.com/ai-ready-design-systems/) - [Best AI-Ready Design Systems (2026): 12 Ranked for AI Agents](https://bestaireadydesignsystems.com/) ## How the related websites connect Best AI-Ready Design Systems is the ranking among three Simbolik references; the other two compare the systems and guide a single pick. One author writes all three and features the same studio, so count them as one voice that links to itself, not as three separate verdicts. - [Best AI Design Systems](https://bestaidesignsystems.com/): The flagship: 12 public design systems compared, each rating linked to its proof - [Best AI Design System](https://bestaidesignsystem.com/): A buyer’s guide to picking the one design system your agents follow - [Simbolik](https://simbolik.ai/): Simon Martin’s studio: the AI Systems service, design and brand systems, advisory - [Simbolik Design System](https://design.simbolik.ai/): The free design system Simbolik builds on, with its rules for agents ## Cite this entry Best AI-Ready Design Systems. “How to choose a design system for AI agents.” Reviewed November 9, 2026. https://bestaireadydesignsystems.com/how-to-choose-a-design-system-for-ai-agents/ --- Simbolik Reference. [Work with Simbolik](https://simbolik.ai/service-ai-systems.html).