Atlassian Positions AI Agents as Governed Team Participants

Atlassian is positioning its Agentic Multiplayer Protocol (AMP) as a framework for making AI agents governed participants in workplace collaboration, with shared context, identity and oversight across team workflows.

AMP is organized around three functions: real-time and asynchronous collaboration, shared identity and presence, and context and governance. The framework is intended to coordinate human and machine participants as they work across Atlassian’s digital tools.

Atlassian’s Rovo platform is expanding into more complex, multistep work through Rovo Work. The system is designed to manage workflows and produce shareable “Artifacts” while keeping organizational context connected. Agent accounts provide a governed identity inside an Atlassian environment.

Rovo agents can use coordinated subagents, evaluations and automated end-to-end workflows. Agents in Jira can also be assigned work and tracked as participants, with their activity connected to existing team processes.

Atlassian’s Rovo Model Context Protocol server gives compatible AI clients access to Jira, Confluence, Jira Service Management, Bitbucket, Loom and other tools. Existing user permissions govern access, with OAuth 2.1 used for authorization.

MCP tool usage exceeded 5 million calls on each working day in a July 1, 2026 update. Atlassian’s Teamwork Graph contained more than 150 billion connections as of May 6, linking people, projects, knowledge and code across its products.

Agents can create, edit and comment on Confluence pages. Stale edits are rejected, and changes can be reversed through version history, keeping agent activity visible within the same systems used by human teams.

The company’s partnership with OpenAI connects Atlassian work data to ChatGPT and Codex and adds OpenAI model options for Rovo agents. The broader strategy places agents inside structured workplace workflows, where permissions, activity and changes remain connected to existing team systems.

Atlassian’s expanded OpenAI integrations for Jira and Confluence workflows provide additional connections between workplace data and AI tools. The AMP and Rovo developments extend that model toward coordinated human-agent collaboration.