Atlassian Introduces AI Agents in Jira to Enable Enterprise-Scale Human–AI Collaboration

Atlassian Corporation has announced the open beta of AI agents in Jira, embedding agentic capabilities directly into the platform where teams plan and track work. 

The move brings Atlassian Rovo and third-party AI agents into core workflows, allowing teams to assign tasks to agents, collaborate with them in comments, and integrate agent-driven execution into everyday business processes—making AI work visible, coordinated, and governed.

Atlassian also unveiled fresh investments in the Model Context Protocol (MCP), reinforcing its strategy to build an open ecosystem where customers can choose and operationalise the right AI agents and tools across their organisations.

“Work is changing fast: people are now orchestrating across agents, tools, and cross-functional teams. Without clear coordination that can easily turn into chaos,” Tamar Yehoshua, Chief Product and AI Officer at Atlassian, said, “We’re focused on helping teams turn that complexity into real productivity.”

With agents in Jira, teams can elevate AI from ad-hoc experiments to accountable teammates. Capabilities include assigning work to Atlassian Rovo and MCP-enabled third-party agents, mentioning agents for in-context iteration, and embedding agents directly into workflows to design, execute, and update work—while humans remain firmly in control. 

Because agents operate within Jira’s existing structures, they adhere to project configurations, permissions, audit trails, and approval flows, enabling confident enterprise adoption.

As enterprises expand the use of agents, MCP provides a consistent standard for agents to access tools, data, and workflows. 

Atlassian reported strong enterprise momentum, with large customers accounting for nearly 50% of all Rovo MCP Server usage, and paid Atlassian editions driving 93% of overall usage.

Building on this adoption, Atlassian announced two major releases. MCP skills in Rovo now let agents connect to MCP-enabled third-party apps—including Amplitude, Box, Canva, Figma, and Intercom—to pull live context and take action. 

Moreover, the Rovo MCP Server has reached general availability, offering an Atlassian-hosted, secure gateway for MCP-compatible AI clients such as Anthropic’s Claude, Cursor, and Google’s Gemini CLI to connect to Jira and Confluence.

Enterprises can begin by trying the open beta of agents in Jira, exploring the MCP server gallery to unlock cross-tool workflows, and extending Atlassian context to compatible AI clients via the Rovo MCP Server—turning AI agents into accountable teammates at scale.