Atlassian turns Jira into mission control for coding agents
Atlassian has introduced new Jira capabilities for assigning, tracking and evaluating work performed by coding agents across enterprise software teams.
Atlassian is repositioning Jira as a control layer for software teams that increasingly work with multiple coding agents rather than one human-only workflow.
What happened
The company introduced Jira capabilities that let teams assign work directly to tools including Claude Code, Cursor and GitHub Copilot, with Codex support planned. Jira can display agent status, connect work to planning and review processes, and automate tasks such as bug fixes and test generation.
Atlassian also wants teams to measure agent output against engineering results, including pull requests and completed work, rather than treating model usage as an isolated cost.
The system uses the company’s Teamwork Graph to provide context from Jira, Confluence, source-code systems and team history. Some capabilities are available to paid Jira Cloud customers, while others remain in early access.
Why it matters
Code generation is becoming easier, but coordinating agents across real projects remains difficult. Companies need permissions, task ownership, status visibility and review checkpoints before agents can safely operate across repositories.
Atlassian already owns the planning and workflow layer used by many software teams, giving it a strong position from which to supervise agent activity.
The bigger picture
The developer-tools market is shifting from AI assistants toward agent operations. The winning platforms may be those that become the system of record for both human and machine work. Atlassian’s move shows that existing enterprise-software vendors can use their workflow data and installed base to compete with newer AI-native tools.