Enterprise Knowledge Search · IT Operations Automation
Atlassian wanted agents to do consequential work inside Jira, Confluence, and related products, but enterprise customers would only trust long-running automation if the models could follow complex instructions, use internal tools safely, and reason across huge contextual graphs.
Reported impact
Atlassian now runs more than 5 million agents in customer business workflows every month, serves millions of users on Rovo Chat, and uses Claude to automate increasingly complex work while keeping context and trust at the center of the platform.
Atlassian deploys Claude on Google Cloud behind the parts of its Rovo platform where the cost of a wrong answer is highest. Rovo Chat uses Claude as one of its base models to choose tools and skills across Jira, Confluence, and the rest of the suite, while Rovo Max takes a goal, pulls context from Teamwork Graph's billions of objects, and works in a managed environment where admins control packages and network access. The same Claude foundation powers Rovo CLI plus code review and generation flows that can read large repositories or pick up Jira tickets directly. Inside Atlassian, engineering runs heavily on Claude Code and recent pilots brought it to product managers and designers for small code changes and prototypes. This architecture lets Atlassian combine broad enterprise context, tool orchestration, and governed execution in one agent stack.
A user goal or Jira/Confluence task enters Rovo Chat or Rovo Max, Claude selects tools and context from Teamwork Graph, works through the task in an admin-controlled environment, and returns actions, code, or artifacts back into Atlassian workflows where humans can keep collaborating.
5 million+ agents executed in customer business workflows each month
Many of Atlassian's customers are also software and IT teams, and Claude handles their code review and code generation
Rovo Max draws on Teamwork Graph, Atlassian's connected map of the billions of objects that span a customer's apps
Someone handed the agent a Jira board and asked it to "create me an Instagram reel"
The Claude models are better at following complex instructions consistently and handling larger amounts of context without losing important details
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
Anthropic Sample Aug 21, 2026 Notion Notion needed to embed trusted Claude agents natively inside collaborative workflows, with automations, long-running tasks, memory, coding sandboxes, and human approvals that could scale across whole workspaces instead… 12 hours of prototyping work collapse into about 20 minutes Claude Managed Agents Anthropic Sample Archived · Apr 16, 2026 Slack Slack needed to turn huge volumes of messages, threads, and files into searchable, permission-aware knowledge and time-saving summaries without sacrificing security or conversational quality. 97 minutes per week saved by the average user through summarization and recap features Claude Code · Claude Platform Anthropic Sample Archived · Feb 10, 2026 HubSpot HubSpot's engineering, marketing, and customer success teams were spending too much time on migrations, maintenance, content creation, troubleshooting, and personalized follow-ups that did not scale across the… From 3-5 days to under an hour Claude Code · Claude Platform · Claude Sonnet 4.5 