Atlassian has launched new Jira and Teamwork Graph options for AI-native software program improvement, geared toward giving engineering groups a single place to handle agent-based work throughout the software program improvement lifecycle.
The adjustments goal a productiveness hole in software program engineering, the place AI device use has risen however beneficial properties in developer pace have remained restricted. Atlassian cited its personal analysis exhibiting a 65% enhance in AI utilization by engineers, whereas developer velocity beneficial properties stayed at about 10%.
The launch centres on Jira as a management level for planning, assigning and monitoring work carried out by human builders and coding brokers. Teamwork Graph offers shared context by linking work, groups, objectives, code and information throughout improvement processes.
In inside benchmarking, brokers utilizing Teamwork Graph context produced outcomes that had been 44% extra correct and used 48% fewer tokens than brokers working with out it.
Planning instruments
Among the many adjustments is Jira for Slack, which lets groups convert Slack conversations into structured specs and work objects utilizing the @Jira immediate. The device can even replace work objects, sync dialogue threads as feedback and assign duties to coding brokers whereas groups proceed collaborating in Slack.
One other addition is Jira Planner, designed for spec-driven improvement. It may possibly pull info from a codebase, Jira historical past, Confluence information and crew context by way of Teamwork Graph to outline necessities and generate a technical specification in Confluence.
Atlassian can also be including Loom video prompts. The characteristic turns a display screen recording and spoken directions into structured process instructions that may be shared with an agent or transformed into Jira work objects.
Agent oversight
Jira can also be being prolonged so customers can assign work on to coding brokers, together with Claude Code, Cursor and GitHub Copilot. Work stays tied to Jira, which acts because the system of file whereas offering context to these instruments.
A built-in Jira Coding Agent is included in each paid plan. It may possibly flip work objects into pull requests prepared for evaluation utilizing Teamwork Graph context and code intelligence, with out requiring builders to arrange a neighborhood atmosphere.
Atlassian can also be introducing a single view for monitoring agent exercise. Engineers can see which agent periods are stalled, which duties are awaiting evaluation and which jobs have been accomplished throughout their workspaces and repositories.
That visibility is a part of a broader push to maneuver coding brokers from restricted trials into wider operational use. Atlassian argues that broader adoption of agent-based work would require governance, automation and measurement instruments inside software program already utilized by engineering groups.
Automation and measurement
New autonomous workflows in Jira enable engineering groups to automate processes utilizing coding brokers inside Jira’s automation rule builder. Groups can ship bug fixes, safety remediation, check era and documentation updates to brokers within the background, then notify engineers when a pull request is prepared for evaluation.
Atlassian can also be launching an Agentic Engineering challenge template and a guided setup wizard to assist groups create tasks with pre-configured workflows, statuses, monitoring and integrations.
For patrons utilizing Atlassian DX, a brand new AI price administration report will mix spending and token information from instruments akin to Claude, Cursor and GitHub Copilot with Jira tasks and crew information. That is supposed to assist groups hyperlink AI spending to engineering outputs and estimate price per pull request.
Sean Joerg, Deputy CISO and Head of Company Engineering at Reddit, described the problem as one in all coordination slightly than mannequin high quality. “The bottleneck in AI-native improvement is not agent functionality, it is coordination at scale to maintain our engineers within the movement,” Joerg stated. “We’re partnering with Atlassian to unravel that: one place the place each agent motion is seen, ruled, and tied to a enterprise end result.”
Trade analysts stated context stays a central concern as extra coding brokers are launched into software program groups. And not using a clear understanding of challenge historical past, technical constraints and inside selections, sooner code era can nonetheless produce poor outcomes.
“As AI coding brokers proliferate, the actual bottleneck is not mannequin intelligence; it is organizational context. Brokers working with out a deep understanding of crew selections, architectural constraints, and challenge historical past produce misaligned code extra shortly, resulting in technical debt and manufacturing points,” stated Jim Mercer, Program Vice President, Software program Improvement, DevOps, and DevSecOps at IDC. “By leveraging Jira and the Teamwork Graph, Atlassian is constructing a context layer for AI. Because the system of file for agile improvement, it could flip tribal information right into a persistent, queryable information layer that may enhance code high quality and launch velocity throughout the enterprise.”
Merchandise now accessible to paid Jira Cloud prospects at no additional price embody brokers in Jira for Claude Code, Cursor and GitHub Copilot, Jira for Slack, the Jira Coding Agent, Jira agent automations, agentic templates and agent periods in Jira. Jira Planner is being supplied by way of a waitlist for early entry, whereas the DX AI price administration report is accessible to Atlassian DX prospects.
“LLMs have made writing code practically immediate. The heavy lifting now could be every little thing round it: defining what to construct, governing what ships, and coordinating throughout people and brokers at scale,” stated Taroon Mandhana, CTO, AI and Teamwork at Atlassian. “Jira has been the system of file for software program groups for 20 years. As we speak, we’re extending that to each agent working alongside them.”