Atlassian announced in September 2026 a set of Jira and DX features for running AI coding agents as continuous, governed loops rather than one-off sessions, citing its 2026 AI SDLC study in which 94% of engineering leaders said they use AI but only 6% have the systems to scale it across the software lifecycle. Most of the features are in open beta or private early access, with others due for general availability in the following months.
"Almost everyone is playing with agents, yet almost no one can let agents run at scale without things breaking," Atlassian wrote on its Inside Atlassian blog.
Atlassian frames the gap as three problems: agents lack knowledge of an organization's architecture, decisions and standards; work runs as isolated sessions that cannot be left running; and leaders cannot measure where AI helps.
For context, Code Context, built on Atlassian's Teamwork Graph, gives Rovo and coding agents secure knowledge of multi-repository codebases for checking backlog ideas for feasibility, producing code-aware implementation plans and triaging bugs, while Agent Context Controls let platform teams decide which agents can work in a space and what they can see. An analysis by DX found that teams whose AI tools used the most Teamwork Graph context shipped roughly 64% more per developer. For execution, agent loops in Jira continuously scan for well-defined, unassigned work items, hand them to Jira Coding Agent to implement and test, and open ready-to-review pull requests in Jira; Standards lets teams define coding standards once and map them to repositories, and AI Review checks every pull request against them before a human sees it. Developers and product managers keep approval of what merges, and the agent updates shared context after work is done. For measurement, DX for Agentic Development tracks AI throughput, quality, adoption and cost, including tool and MCP tracking and model-to-task fit, and a Jira Agent Usage Dashboard shows which agents teams use.
Code Context is rolling out in open beta to paid customers; agent loops, Standards and AI Review are in private early access; Agent Context Controls and the Agent Usage Dashboard are due for general availability in the coming months, and DX for Agentic Development within the quarter.
Details
- City
- Sydney
- Organization
- Atlassian