Atlassian opens up the 'Team Collaboration Map' to external parties... Robo AI will also expand into an intelligent agent

Atlassian publicly announced major AI updates at its annual event “Team ’26.” The core is opening up the “Team Collaboration Graph” that connects workflows within organizations, and expanding its AI assistant “Rovo” into an “agent” capable of autonomous planning and executing multiple steps.

Atlassian explained that the team collaboration graph, as a “shared context layer,” links personnel, projects, documents, and decision data to the Atlassian product suite and external tools. According to the company, the number of connections within the graph has now exceeded 150 billion. The two features open for testing are: a command-line interface (CLI) for the collaboration graph aimed at developers; and a tool that enables graph interoperability via Rovo’s Model Context Protocol (MCP) server.

The new CLI supports over 300 commands. With it, coding agents like Anthropic’s Claude Code or Cursor can query Atlassian’s entire workflow network through a single interface without stitching together APIs for each product. MCP interoperability aims to allow external AI tools following this standard to read and write data within the collaboration graph.

Atlassian stated that in its internal benchmarks, generating AI responses based on the collaboration graph data improved accuracy by 44% and reduced token usage by 48%. For enterprises, this means better response quality and cost efficiency. Additionally, a collaboration graph connector based on the Forge platform has been officially released, enabling customers to connect their own or legacy systems’ data to the graph while maintaining existing permission structures.

Rovo: Beyond “Conversational AI,” Moving Toward Actual Task Execution

Rovo’s evolution is also notable. According to Atlassian, over 14 million Rovo-supported tasks have been executed by customers in the past month. The overall intelligent agent automation platform has grown sevenfold in the past six months. The company added that over 90% of enterprise cloud customers are now using Rovo.

The upcoming “Max” reasoning mode, available in early access, decomposes complex requests into multi-step plans, executes them across interconnected tools, and returns intermediate results for user review. This marks a shift from simple Q&A to “executable AI.”

A no-code environment called “Rovo Studio” for creating agents has also been officially launched. It includes role configuration, approval workflows, version control, and audit features, enhancing operational stability for enterprise use. As generative AI applications increase, security and control issues are becoming more prominent, and Atlassian appears to be focusing on “operational AI” rather than just productivity.

AI Expansion into Jira, Service Management, and Engineering Tools

The scope of AI applications across all products is expanding. The intelligent agent feature in Jira has been officially released, capable of directly receiving and executing work tasks with audit logs kept. Jira Product Discovery Enterprise has been launched with hierarchical governance features, and a new “Feedback” feature for collecting customer signals is in early access.

The newly announced “Incident Command Center” is an integrated product for fault detection, investigation, and resolution. It includes root cause analysis powered by Rovo, while “Rovo Service” offers autonomous or supervised Tier 1 customer support. Its structure involves AI handling repetitive initial response tasks.

A browser-based reporting tool, “Dia Reports,” has also been released. It combines team collaboration graph context with daily work tools to proactively generate customized briefs such as interview prep documents or decision memos. It works by prompting the user for input after presenting the required report, shifting the AI interface from a “question-answer window” to a “pre-structured work layer.”

Features for engineering organizations have also increased. New features include: “Agent Experience,” which measures codebase interactions with agents; “AI Code Insights,” which tracks AI-generated code at the commit level; and “AI Pulse,” a productivity signal tool for development managers.

Core of AI Diffusion: “Context” and Governance

The most important aspect of this release is Atlassian’s focus on “organizational context” rather than just “model performance” in AI competition. Its strategy is not merely to create smarter chatbots but to integrate personnel, documents, workflows, and decision history into a unified organizational fabric, enabling AI to operate within actual company operational structures.

Management functions are also enhanced accordingly. A new organizational-level intelligent agent list can display in real time who created which agents, where they run, and their usage frequency. AI access permissions are separated from agent creation rights to prevent uncontrolled proliferation. Dashboards and audit logs track AI usage and token consumption. Additionally, strategies for external data import, data storage locations, and control over large language models hosted by Atlassian are provided.

Matthew Hargreaves, Head of Product Delivery and Automation at Rendi Group, commented: “Rovo and Atlassian’s team collaboration graph are the core links connecting Jira, Confluence, JSM, Slack, email, and others. This is a turning point, marking AI’s shift from peripheral tools into the core of organizational operations.”

Atlassian CTO Andrew Boyagi also emphasized in a recent interview that organizational context is essential for AI agents to generate real value. This aligns with recent enterprise AI market trends. As general model competition intensifies, the depth of understanding of our data and workflows is becoming a key success factor.

This release clearly demonstrates that trend. AI applications have moved beyond assisting work tools, into a comprehensive competition to enable AI to understand and execute organizational operational systems themselves.