Atlassian is turning its Teamwork Graph into a broader data layer for AI agents and developer tools. The company launched a developer command-line interface and Rovo Model Context Protocol server in open beta, allowing external assistants including Claude Code, Cursor, and ChatGPT to interact with Atlassian’s work data instead of relying on separate product-by-product integrations.
The expansion coincides with the general availability of Teamwork Graph Connectors via Forge, Atlassian’s cloud development platform. This deployment allows enterprise corporate partners to integrate proprietary, industry-specific data systems into the graph while preserving existing security access controls. According to internal benchmarks, grounding AI responses in Teamwork Graph data delivers 44% more accurate results and uses 48% fewer tokens compared to standalone product searches. The company also rolled out an independent destination, TeamworkGraph.com, where users can view collaboration patterns and receive automated data source recommendations without writing code.
The updated data infrastructure has already been deployed across specialized corporate workflows to streamline product operations. By constructing custom Forge connectors for automotive defect management and engineering traceability lines, Mercedes-Benz linked test cases, components, and vehicle models into the central data ecosystem. Magnus Östberg, Chief Software Officer at Mercedes-Benz, stated that “having agents find, connect, and act on information from across our Atlassian platform is a great enabler.”