Atlassian is shipping four governed agent loop features across Jira and DX that automate the path from backlog to pull request, backed by its 2026 AI SDLC study showing 94% of engineering leaders use AI but only 6% can scale it across the software lifecycle.
"You stay in control of the merge button but stop being the bottleneck for everything leading up to it," wrote Taroon Mandhana on Atlassian's engineering blog on September 10, 2026.
Most AI in engineering today is one prompt at a time — individual magic moments that don't scale to orgs of hundreds of engineers. Agents fail without context about a team's architecture, decisions, standards, or institutional memory, and without governance the gap between a demo and something an org can trust stays wide. Atlassian frames the answer as moving from ad hoc agent sessions to always-on workflows.
Atlassian shipped Code Context on its Teamwork Graph as the foundation, then layered governance via Agent Context Controls that let platform teams constrain which agents operate in a space and what they can see. The next phase turns backlogs into always-on automated execution cycles via Agent loops in Jira, Standards for shared coding standards across repositories, and AI Review putting a dedicated agent on every pull request. Visibility and accountability arrive via DX for Agentic Development and Jira Agent Usage Dashboard.
The architecture has three layers. Context: Code Context on Teamwork Graph gives Rovo and coding agents secure intelligence across multi-repo codebases; Agent Context Controls govern which agents operate and what they see. Execution: Agent loops in Jira continuously scan for unassigned work items, delegate to Jira Coding Agent for execution and testing, and open ready-to-review PRs directly in Jira; Standards maps organizational coding standards to repositories; AI Review checks every PR against those standards. Measurement: DX for Agentic Development unifies AI Code Insights, tool and MCP tracking, model-to-task fit, and Agent Experience research; Jira Agent Usage Dashboard surfaces which agents are used.
DX analysis of teams whose AI tools used the most Atlassian Teamwork Graph context found they shipped roughly 64% more per developer. The 2026 AI SDLC survey found 94% are using AI but only 6% have systems to scale it across the lifecycle. Code Context is rolling out to paid Atlassian customers through open beta; Agent loops, Standards, and AI Review are in private early access; Agent Context Controls and the Agent Usage Dashboard are slated for general availability in the coming months; DX for Agentic Development is generally available for Atlassian DX customers this quarter.
Atlassian is hosting a State of AI SDLC digital summit on September 22, 2026, framing governed agent loops as the path from individual magic moments to a system an engineering org can run on: developers define intent and guardrails, agents execute in parallel, and humans retain approval over what ships.