Asana builds human-agent teams using Claude

Asana integrates Claude-powered AI agents directly into its Work Graph model, giving them defined roles, credentials, and access controls alongside human collaborators. Employees use Claude to turn unstructured ideas and conversations into structured tasks, which agents then act on. Each agent has a profile listing its purpose, permissions, integrations, and administrators. Access is intentionally scoped, and an agent's effective permissions are bounded by whoever triggers it. Agents also retain shared memory across interactions, allowing teams to coach and reuse that knowledge, though only designated administrators can modify how an agent behaves.