Atlassian today has unveiled a massive suite of AI-driven updates at its annual Team event, signaling a major shift toward an AI-native ecosystem. Central to the announcement is the expansion of the Teamwork Graph and the evolution of Rovo, Atlassian’s flagship AI agent, which are designed to harness organizational context to deliver unprecedented productivity gains.

The Teamwork Graph: A Unified Source of Truth

The bedrock of Atlassian’s AI strategy is the Teamwork Graph, a context layer that maps over 150 billion objects and relationships across an organization’s projects, goals, and teams. 

Unlike traditional AI tools that rely on superficial searches and often “stuff” context windows with irrelevant data, the Teamwork Graph allows AI to traverse deep relationships. Atlassian reports that this approach makes AI answers 44% more accurate while using 48% fewer tokens.

To empower developers, Atlassian is releasing a Teamwork Graph CLI and MCP tools. These allow customers to plug their organizational memory into external AI builder tools like Cursor or Claude, ensuring that custom-built apps benefit from the same rich context available within Jira and Confluence.

“We already have over 150 billion objects and relationships mapped in the Teamwork Graph, and billions of those are changing every month,” Jamil Valliani, who runs the AI product team at Atlassian, told SD Times. “That is a rich source of truth for us and for our Rovo and other applications to use to deliver increasingly better quality results and capabilities and tools to our customers.”

Rovo Max: Pushing the Limits of AI Agency

The spotlight also fell on Rovo, Atlassian’s AI-powered tool for searching organizational data, which has seen a 50% increase in usage quarter-over-quarter, the company said. 

Atlassian today introduced Rovo Max, a new mode for solving high-complexity problems. When activated, Rovo Max spins up a virtual machine in the cloud to decompose tasks, write code for analysis, and test itself autonomously. In one demonstration, Rovo Max generated a professional-quality audio podcast summarizing multiple Confluence pages—a skill it was never explicitly taught, but learned on the fly by researching best practices.

Bridging the Gap Between Search and Action

Beyond agents, Atlassian is revolutionizing enterprise search. New connectors now link Rovo to over 50 third-party applications, including SharePoint, Slack, and Salesforce. This integration allows Jira users to reference Google Drive documents or Salesforce records directly within their workflow. “We can provide high-quality search across all of those different applications,” Valliani said.

In Jira, agents are becoming first-class citizens. Users can now assign issues to AI agents as easily as to humans. These agents can capture plans in Jira, update progress, and dialogue with team members through comments. Early data shows a 7x increase in agent-led automations, with companies like Mercedes-Benz already using these tools to improve bug report quality.

Valliani said working on search is important so customers can more easily find the documents that matter to them. “We make sure that we find the right bits of information for the customer, both through graph and through search, to power the scenarios that matter most to them and use the best possible data.” That data, he said, is not just from Atlassian objects but includes data from the rest of the organization.  “We’re able to deliver profoundly better experiences and accelerate workflows just by having that capability,” he said. 

Studio Updates

The Agent Building Studio has received significant investment to enhance its capabilities beyond simple agent creation, now supporting automation and the development of whole new apps. The goal is to make AI building more accessible, even for users without sophisticated technical knowledge. 

Valliani said, “Users can now simply use natural language prompts, such as, ‘I need to automate X, Y, Z, so that every day at 9am X, Y and Z happen.” The Studio automatically generates a plan, including necessary agents and automations, for review. It has improved testing capabilities, allowing users to dialogue with and test the newly created agent in a panel. If there is an issue, users can simply describe the problem, and Studio will correct it, rather than requiring the user to update the underlying prompt. 

The Future of the SDLC and Developer Joy

For software teams, Code Intelligence now offers semantic understanding of massive codebases. Rovo can reason over 20 years of code in minutes to identify style inconsistencies or security vulnerabilities. Looking ahead, Atlassian teased “Rovo Dev,” which will soon be capable of autonomously submitting Pull Requests (PRs).

Reflecting on the joy of development, Atlassian emphasized that these tools are not meant to replace creativity but to remove the “tedious machinery” of status reports and compliance. By delegating the nitty-gritty to AI, developers are free to focus on solving complex problems and building products that Valliani said “make the customer smile.”