Atlassian has recently taken a significant step by opening its Teamwork Graph to third-party AI agents and tools, thereby broadening its data layer to include external applications.
This graph encompasses over 150 billion connections that link people, tasks, objectives, code, and content from Atlassian’s own software and integrated applications. Through new open beta tools, Atlassian is enabling external AI systems to search, reason, and act using this contextual data, all while adhering to existing permissions and administrative controls.
This initiative marks a notable change in how Atlassian envisions the use of its AI products. Instead of restrictively confining its Rovo assistant within a closed environment, Atlassian is positioning its foundational graph as a communal resource for a broader array of AI clients, coding tools, and workflow applications.
Mike Cannon-Brookes, the company’s CEO and Co-founder, articulated this vision in a statement accompanying the launch. “In 2026, anyone can buy ‘smarts’ by the token,” he remarked.
“The true advantage lies in your institutional memory: every plan, document, and decision that your teams have ever made. Rovo serves as the interface that transforms intelligence and context into genuine progress for your organization.”
Open Access
One of the new ways to access the Teamwork Graph is via an MCP server now in open beta. This server will enable clients compatible with the Model Context Protocol, including tools like Figma and Replit, to interact with the graph through a standardized interface.
A second option is a command-line interface, also in open beta, targeted at software developers. Coding agents such as Anthropic’s Claude Code and OpenAI’s Codex will have the ability to directly utilize graph data within their development environments, while administrators maintain control over access permissions and scopes.
This strategy showcases a wider competition among software firms regarding the placement of business context as AI agents take on a more prominent role in office tasks and software development. By providing standardized access points to its graph, Atlassian is betting that the governance of structured workplace context will be more crucial than the AI models themselves.
Atlassian linked this launch to the increasing adoption of Rovo by major clients. Currently, Rovo is utilized by over 75% of Fortune 500 companies and 90% of its enterprise cloud users.
In the last month alone, customers executed more than 14 million actions assisted by Rovo, and automations powered by agents have surged sevenfold over the past six months.
Product Changes
In addition to opening the graph, Atlassian unveiled an extensive suite of AI-related product updates across its Jira, Confluence, developer tools, and product management software. These updates aim to transition Rovo from a mere assistant responding to prompts into capable software agents that can be assigned specific tasks or handle multi-step processes.
A new Rovo Studio workspace is now generally available, providing users with a centralized platform to create agents, automate tasks, and develop custom applications using natural language. The inclusion of governance features is crucial for larger clients that need to regulate how AI tools interact with their internal systems and data.
Atlassian also introduced Max mode for Rovo Chat, a feature set to launch soon. This functionality will take complex requests, break them down into actionable plans, and execute tasks across various Atlassian products and connected services.
In Jira, agents have now become generally available. Teams can assign tasks to these agents, mention them in comments, and integrate them into workflows and automations, thereby giving AI tools a more explicit role in project management and execution.
Confluence is set to receive updates including a beta feature called Remix with Rovo, enabling users to convert content blocks into formats like charts, databases, and infographics directly on the page. Additionally, a slides feature for Confluence will follow in beta, allowing users to create presentation materials while keeping the original content intact.
For software development teams, Atlassian’s DX AI Experience is now generally available. This feature aims to assist engineering organizations in tracking AI-generated code, evaluating agent performance, and measuring the return on investment from AI usage throughout the software development lifecycle.
The company is also launching an early access Product Collection, which consolidates Jira Product Discovery, the Feedback app, and other elements, including an anticipated Pendo integration. This collection aims to provide product teams with a unified environment for capturing feedback, ranking ideas, and managing delivery tasks.
Additionally, Atlassian has introduced Dia, its AI browser, which is now available for teams to explore. Dia is designed to generate daily briefings from a mix of emails, messages, calendars, graph context, and browsing activity and supports single sign-on and mobile device management for Chromium, along with SOC 2 Type II attestation.
In summary, these updates represent Atlassian’s bold push into the workplace AI software market by linking agents to the records, workflows, and documents within its suite of collaboration tools. The fundamental belief driving this strategy is that AI systems will be far more beneficial when they can tap into an organization’s historical context of plans, discussions, and decisions.
