Atlassian's use AI to design and build every day, but AI tools are only as good as the context they have. We translated the Atlassian Design System’s documentation content into machine-readable schema files to deliver context to agents using ADS MCP, Skills and more to improve their accuracy and generation speed while reducing AI token costs.

Atlassian provides insights into software development collaboration, project management, and team productivity, offering documentation, tutorials, and best practices for using Atlassian's suite of tools, including Jira, Confluence, and Bitbucket. By exploring Atlassian's curated content, developers can learn about agile methodologies, version control workflows, and continuous integration practices for managing software development projects effectively. Whether you're working in a small startup or a large enterprise, Atlassian offers resources to streamline your collaboration and drive project success.

Atlassian

Atlassian's design system team tackled the problem of AI tools generating inaccurate UI by restructuring their Atlassian Design System (ADS) documentation into machine-readable schema files. These TypeScript-based schemas cover components, tokens, icons, lint rules, and accessibility requirements. The structured content powers an ADS MCP server, design system skills, and DESIGN.md files. Compared to agents working without an MCP, the approach yielded up to 52% accuracy improvement on specific queries, 34% faster task completion, 16% fewer tokens consumed, and 26% fewer tool calls. Self-healing tooling keeps documentation in sync with the codebase automatically, and the team plans to connect structured content to more agents and the Teamwork Graph.

Teaching AI to speak our design language

What structured content looks like in our design system

Early results show more accurate and faster UI generation

How structured content shows up in real workflows