Atlassian upgrades AI coding agents for always-on software development

Atlassian plc has unveiled new Jira features designed to empower engineering teams to deploy AI agents at scale, managing their operations continuously throughout software development. As organizations increasingly adopt these agents, Atlassian emphasizes the importance of extended operational periods and broader lifecycle integration.

However, this approach brings challenges related to trust, grounding, shared context, communication, institutional memory, and validation. To tackle these issues, the company is rolling out updates to control agentic behavior, establish action standards, periodically review AI agents in operation, and ensure proper usage. Atlassian addresses a common failure point for agents: insufficient understanding of project architectures, decisions, and standards.

To combat this, the company introduces "Code Context," bolstered by its Teamwork Graph, providing secure contextual intelligence across intricate, multi-repository codebases. This is complemented by "Agent Space Settings" and "Agent Context Controls," enabling teams to securely manage agent operations, akin to employee access controls.

The ultimate goal is to move beyond traditional chatbot functionality, enabling AI agents to operate independently, executing tasks over extended periods with minimal oversight. Key innovations include AutoDev, which automatically transforms backlogs into merge requests within Jira; a standards system ensuring code quality; an agent dedicated to scrutinizing merge requests against established guidelines; and a dedicated agent for generating or updating technical documentation in Confluence from code repositories.

Throughout this process, transparency is prioritized through the generation of audit logs for AI agent activities, which include diagnostics to measure impact on throughput, quality, adoption, and cost. Additionally, an AI agent usage dashboard provides insights into tool utilization, user behavior, and outcomes at the team level.

By offering these accountability and visibility tools, Atlassian facilitates a structured approach to integrating AI agents into engineering workflows, transitioning teams from sporadic ad-hoc experimentation to a more systematic and efficient use of always-on agents that handle backend tasks to alleviate engineering burdens.