How Asana Uses Claude to Build AI Teammates That Cut Review Cycles to 15 Min

Asana is a work management platform used by enterprises to coordinate projects and cross-functional workflows. As AI capabilities matured, Asana saw an opportunity to embed autonomous agents directly into the collaboration layer that teams already used daily, rather than adding a separate AI tool that required context-switching.

The challenge was building agents that could take meaningful, multi-step actions while maintaining the auditability and control that enterprise customers require. Agents needed to work within existing Asana task structures, pull from the right data sources, and hand off to humans at the right moments — without becoming a black box.

Asana built AI Teammates on Claude Opus 4.6 using a multi-agent architecture. Claude Opus 4.6 handles top-level orchestration, while specialized subagents powered by different Claude models run specific tasks optimized for either speed or reasoning depth. Teammates integrate with Asana's Work Graph — the structured representation of goals, projects, tasks, and dependencies — giving agents the context they need to act accurately. All agent actions are logged in task history for full auditability, and administrators control access and scope.

The results are measurable in time saved on high-volume knowledge work. Campaign brief review cycles that previously spanned days now complete in approximately 15 minutes with the Campaign Brief Writer teammate. The Launch Planner monitors cross-functional dependencies and surfaces blockers proactively, eliminating manual spreadsheet tracking for program managers. HR Request Triage routes employee requests without manual handling. Compliance Reviewer evaluates submissions against regulatory requirements autonomously.