Atlassian wants developers to finally like Jira

Atlassian has a developer problem, and it knows it. "It's known, developers don't like to interact with Jira," says Ming Wu, the company's head of engineering for DevAI — an admission that could double as the top comment on any Hacker News thread about the tool. On Wednesday, Atlassian rolled out a batch of AI features meant to change that calculus, betting that if Jira can host the actual work instead of just tracking it, developers might stick around.

The centerpiece is the Jira Coding Agent, previously known internally as Rovo Dev, now bundled into every paid plan. It runs in the cloud rather than on a laptop, turning work items straight into pull requests without anyone spinning up a local environment. Every ticket now carries a choice: assign it to a human, or to an agent, sitting side by side. Teams can also route work to Claude Code, Cursor, or GitHub Copilot, with OpenAI's Codex expected later. A new Jira Planner draws on Jira and Confluence history to draft a technical spec in Confluence, and Jira for Slack lets anyone file a ticket just by tagging @Jira in a thread. Automation rules can quietly hand off bug fixes, vulnerability patches, and test generation to agents running in the background.

Wu argues the real bottleneck was never writing code — it's the human doing the reviewing, and specifically the mental switching cost of jumping back into someone else's context. She points to guided tours and AI auto-review as ways to narrow what a person actually has to look at, though she's blunt that long code stays hard to review no matter how much tooling surrounds it. She goes further, predicting that editors themselves may eventually fade — "the more agents write code, the less the developer will write code" — while insisting that doesn't erase the need for product managers or engineering leads, so much as blur the lines between what each role does.

None of this is happening without scrutiny. Atlassian ran a longitudinal study with DX, the developer-productivity platform it acquired for $1 billion last year, and found AI-driven productivity gains topping out around 15%, with plenty of teams landing below 10% — a far cry from the mythical 10x. Atlassian's own internal test, run across 6,000 of its engineers using the new tools, reported a 44% jump in agent task completion efficiency, a 48% drop in token consumption, a 36% cut in PR cycle time, and just over half of routine code vulnerabilities resolved automatically and queued up for developer review. Wu credits much of that to the Teamwork Graph, Atlassian's context layer spanning Jira, Confluence, Slack, GitHub, and Jira Product Discovery, though she's careful to flag the usual caveats around sample size and bias in any vendor's numbers, including her own.

Atlassian isn't pretending to have this fully mapped out. Wu says the company plans roughly a quarter ahead, with six months as the practical ceiling before it's just guessing. Given how fast the coding-agent space is moving, that might be the most honest thing anyone at Atlassian says all week.