Atlassian Rovo

Atlassian Rovo is the AI layer across Atlassian's cloud products, providing search, chat, autonomous agents and the Rovo Studio builder inside Confluence, Jira, Jira Service Management, Bitbucket, Trello and Loom. It serves engineering, product, IT service and operations teams that already work in Atlassian and want AI that carries project context between a ticket, the spec that spawned it and the pull request that implements it. Rovo is now also sold as its own app: the Rovo Standard user role gives people access without a paid Jira or Confluence subscription. Rovo is built on the Teamwork Graph, a contextual layer connecting people, projects, goals, decisions and code, so an answer can carry context from a Jira issue to the Confluence spec behind it and the Bitbucket pull request that implements it. Atlassian publishes around twenty third-party connectors on the Rovo page, including Google Drive, Gmail, Google Calendar, Microsoft Teams, SharePoint, OneDrive, Outlook Mail, GitHub, GitLab, Azure DevOps, Box, Dropbox, Figma, Miro, Lucid, Smartsheet, Docusign and Zendesk. Rovo runs Model Context Protocol in both directions. Atlassian operates an official remote MCP server, so external clients such as Claude, ChatGPT and Cursor can read Atlassian data, and Rovo itself ships nine out-of-the-box agents that consume third-party MCP servers inside Confluence and Jira, among them Intercom and Lovable, with Figma, Canva, Box, Amplitude, HubSpot and Gamma generally available and Replit in beta. Pricing is in USD and runs on two ladders. Rovo Standard is charged at $5 per user per month per organization and is currently in open beta, so it is not billed yet. When the beta ends an organization returns to Rovo Free unless it opts in to the paid plan, with at least 90 days to decide. Rovo is otherwise included with paid Atlassian Cloud plans rather than sold separately. Confluence Standard is $6.70 per user per month and Premium $13.20 at one to one hundred users, both falling with seat volume, and Enterprise is quote-only and requires 801 or more users. What the entry tier does not buy: Rovo is metered in credits, and the allowance rather than the seat price governs how much AI a team gets. Standard includes 25 Rovo credits and 100 indexed objects per user per month, Premium 70 credits and 250 objects, and Enterprise 150 credits and 625 objects. Extra usage is disabled by default and an admin must switch it on, with alerts at 80 and 100 percent of the monthly allowance. The nine third-party MCP agents require Rovo Premium or Enterprise, and a site admin must enable each agent's MCP server. HIPAA coverage is scoped: Rovo Chat, Rovo Agents and Rovo Search are supported in HIPAA-enabled environments, while the Chrome extension, the CLI, Rovo MCP, the mobile app and the desktop app are out of scope for protected health information. Rovo is not available on Atlassian Government Cloud. Rovo is built for teams with an Atlassian footprint, because even the standalone Rovo Standard role sits inside an Atlassian organization. Teams that want an AI workspace with its own documents and enterprise search should look at Notion AI at $10/mo. Teams whose work already lives in Word, Excel, PowerPoint, Outlook and Teams will find equivalent coverage in Microsoft 365 Copilot at $30/mo. Teams that want the same idea inside a project-management platform rather than a wiki should look at ClickUp Brain at $7/mo billed annually. Teams that want their own on-screen and meeting activity captured locally on the device, rather than search over work already filed in cloud tools, should look at screenpipe. Current state Q3 2026: Rovo Standard is in open beta and free during it, with a published rate of $5 per user per month per organization to follow. Rovo holds SOC 2 Type II and ISO 27001 attestations of its own, alongside GDPR commitments and the scoped HIPAA support above. Third-party LLM providers, including OpenAI, Anthropic and Google, operate under zero data retention agreements and do not train on customer inputs or outputs. Atlassian may fine-tune its own open-source models on de-identified, aggregated metadata, subject to admin-controlled data contribution settings. Named customers include Mercedes-Benz, Cisco, Reddit, Procore, Sprout Social, Datasite, Harper Collins Publishers and the Atlassian Williams F1 Team.