Atlassian launches 'always-on' AI agents to automate Jira backlogs

Atlassian has introduced a suite of "always-on" capabilities for its AI coding agents across Jira, Confluence, and DX. The update marks a strategic pivot from isolated AI prompts toward integrated, governed workflows that connect agent context, execution, and measurement.

At the center of the release are "agentic loops" within Jira. These loops continuously scan for well-defined, unassigned work items in the backlog, delegate them to the Jira Coding Agent for execution and testing, and automatically open pull requests (PRs) directly in Jira. To support this autonomy, Atlassian introduced "Code Context," a feature built on the company's Teamwork Graph that provides Rovo and other coding agents with secure intelligence across multi-repository codebases.

Solving the Context Gap

Atlassian is positioning its existing tools as the "system of record" for team collaboration to solve a persistent problem in enterprise AI: the context gap. While standalone AI assistants often lack the broader organizational knowledge required for complex tasks, these integrated agents leverage the planning and delivery data already present in Jira and Confluence.

By orchestrating agents alongside human engineers, Atlassian aims to provide teams with a safe, measurable way to scale agentic workflows across the software development lifecycle (SDLC). The goal is to address the primary bottleneck in AI software engineering: organizational context rather than model intelligence.

Governance and Measurement

To address the security and accountability concerns inherent in autonomous AI, Atlassian has added several governance layers. Agent Context Controls allow platform teams to manage which agents can operate in specific spaces and restrict what those agents are permitted to see. Additionally, a new Jira Agent Usage Dashboard enables team leaders to track agent sessions and monitor delivery velocity.

Quality control is further managed through a "Standards" feature. This allows platform teams to define specific organizational coding standards, which a dedicated AI Review agent then uses to flag issues in PRs before the code is shipped.

Tracking AI ROI

To quantify the impact of these autonomous workflows, Atlassian introduced DX for Agentic Development. This measurement tool tracks AI performance across four key metrics: throughput, quality, adoption, and cost. The system unifies AI Code Insights, tool and MCP tracking, and Agent Experience (AX) research into a single view.

Industry observers will now be watching how enterprises adopt these autonomous loops in production. While the tools provide the necessary governance and measurement, the transition from human-prompted AI to AI that autonomously monitors and executes a backlog represents a significant shift in the traditional software development lifecycle.