Atlassian has been quietly running the same play since 2023: instead of selling AI as an add-on, embed it so deeply into Jira and Confluence that it becomes invisible infrastructure. Rovo is that infrastructure — and after Team ‘26 in May 2026, it’s no longer accurate to call it an AI assistant. It’s an agentic platform.

More than 90% of Atlassian’s enterprise cloud customers are now using Rovo. That number matters less as a testimonial and more as a signal: Rovo is not optional for teams in the Atlassian ecosystem. It’s already there. The question is whether teams are using it well.

Category: AI Assistants

At a Glance
Product Atlassian Rovo
Type AI platform (search + chat + agents + dev tooling)
Bundled with All paid Atlassian plans (Free, Standard, Premium, Enterprise)
Context source Teamwork Graph — 150B+ connections across Jira, Confluence, Bitbucket, and 50+ integrations
Monthly actions 14M+ Rovo-assisted actions (as of May 2026)
Agentic growth 7x increase in agentic automations over 6 months
Enterprise adoption 90%+ of enterprise cloud customers
Rovo Dev Separate — $20/developer/month, 2,000 monthly credits

What Rovo Is

Rovo is the product umbrella for all AI features across Atlassian’s cloud products. It has six components, each solving a different layer of the “teams don’t have time to do everything” problem.

The underlying engine is the Teamwork Graph — Atlassian’s internal knowledge graph with over 150 billion connections: people, projects, Jira tickets, Confluence pages, Bitbucket PRs, OKRs, incidents, and decisions. The graph is what makes Rovo contextually useful in ways a generic AI assistant can’t replicate. A question like “what’s blocking the payments redesign?” isn’t answerable from a document — it requires traversing relationships across teams, sprints, and dependencies. The Teamwork Graph is how Rovo answers it.

The Six Components

Cross-tool search across Jira, Confluence, and 50+ connected apps (Google Drive, SharePoint, Slack, GitHub, and more). Rather than searching within one tool at a time, Rovo Search finds the answer across your entire organizational knowledge base in a single query.

The practical difference from standard Atlassian search: Rovo Search understands intent, not just keywords. “Find the architecture decision we made about the payment service” works even if no document is titled that.

2. Rovo Chat

A conversational interface grounded in your company’s actual context. Unlike generic AI chat (ChatGPT, Claude), Rovo Chat can:

  • Summarize the last two weeks of a Jira project
  • Draft a Confluence page based on a set of tickets
  • Answer “who owns this?” using Teamwork Graph relationships
  • Explain what a team has been working on without you having to find all the relevant docs

Rovo Chat is available in Jira, Confluence, and the standalone Rovo web interface. As of Team ‘26, the Max mode (reasoning mode) is entering early access. Max builds a multi-step plan rather than a single reply — querying Jira statuses, pulling Confluence decisions, checking support queues — and loops you in at review points. It’s the upgrade from “chat that helps” to “agent that executes.”

3. Rovo Agents

Pre-built and custom agents that execute work inside Jira and Confluence autonomously. GA as of Team ‘26, with full audit logging.

What agents can do:

  • Jira agents: Assigned to work items, they triage backlogs, update issues, create tickets from conversation context, and move tickets through workflows
  • Confluence agents: Summarize meeting notes into action items, maintain documentation, update pages when linked Jira tickets resolve

Agents in Jira can now be assigned work items the same way you’d assign a human team member. The audit trail is complete — every action an agent takes is logged.

4. Rovo Studio

A no-code/low-code builder for creating custom Rovo Agents and automations. GA as of Team ‘26, with added enterprise controls: built-in roles, approvals, versioning, and audit controls.

Studio agents are grounded in the Teamwork Graph by default — you’re not building a generic bot, you’re building an agent that understands your org’s project structure, team relationships, and tool stack. This is meaningful for organizations that have specific workflows that don’t fit the pre-built agent templates.

5. Rovo Dev

Rovo Dev is the developer-facing surface of Rovo. It is sold separately ($20/developer/month, 2,000 monthly Rovo Dev credits) and has three main capabilities:

Code generation from Jira context: Rovo Dev reads the issue, creates a code plan, generates code, and can open a Bitbucket PR automatically. The loop from ticket to PR runs without context-switching.

AI code review: Available in both Bitbucket and GitHub. Rovo Dev reviews PRs against the acceptance criteria in the linked Jira ticket — not just syntactic review, but semantic validation against the original requirement. Atlassian measured a 30.8% reduction in PR cycle time in their internal rollout across 80,000 developers.

VS Code integration: Rovo Dev is GA in VS Code (and compatible with Cursor, Windsurf, and other VS Code-based editors). It provides IDE-native search, chat, code editing, test running, and commit preparation, with Jira and Confluence context visible in the sidebar.

6. Rovo Max (Early Access)

Announced at Team ‘26. Max is a reasoning mode for Rovo Chat that changes the interaction model: instead of replying to a single prompt, it builds a plan, executes it across connected tools, and presents results for review.

A Max session might:

  1. Pull open Jira issues for a sprint
  2. Cross-reference blockers in Confluence
  3. Check recent Slack/support signals (via integrations)
  4. Draft a status summary and proposed next actions
  5. Offer to create follow-up tickets

Max is not yet GA — it’s in early access as of May 2026.

Pricing

Rovo Search, Chat, and Agents are bundled into every paid Atlassian plan at no additional line item. Credits are included per seat:

Plan Monthly Rovo Credits/User
Standard 25
Premium 70
Enterprise 150

As of May 2026, Atlassian is not billing for overages — and has committed to at least 90 days’ notice before changing that, with an explicit opt-in requirement before metered billing begins.

Rovo Dev is priced separately: $20/developer/month with 2,000 monthly Rovo Dev credits. Overages are billed at published rates.

After Team ‘26: The Strategic Shift

The most important thing to understand about Rovo is the framing shift from Team ‘26. Atlassian is no longer positioning Rovo as “AI features in your tools.” The new framing is: Atlassian products are the execution layer for enterprise AI agents.

This has two implications:

  1. Rovo agents are first-class team members — they can be assigned work items, they participate in workflows, they leave audit trails. This is not a sidebar assistant; it’s an actor in your processes.

  2. The Teamwork Graph is open to external agents — via the Rovo MCP Server and a new Teamwork Graph CLI. Any MCP-compatible agent (Claude Code, Cursor, Codex, etc.) can now query Atlassian’s org context. See our Atlassian Teamwork Graph builders-log for the strategic analysis.

Who Rovo Is For

Strong fit:

  • Teams already in the Atlassian stack — Rovo’s advantage is grounded context; it doesn’t deliver that value for teams not using Jira/Confluence heavily
  • Enterprise teams with high documentation and ticket volume — the Teamwork Graph gets more valuable as the data grows
  • Engineering teams evaluating AI coding tools — Rovo Dev’s Jira-to-PR loop is a meaningful workflow integration that GitHub Copilot and Cursor don’t replicate

Weaker fit:

  • Small teams with light Atlassian usage — the credit bundle is there, but there isn’t enough graph data to make Rovo’s context advantages meaningful
  • Teams primarily on non-Atlassian tools — Rovo integrates with Google Drive, Slack, GitHub, and others, but the native advantage disappears as the Atlassian footprint shrinks
  • Teams needing Max mode today — it’s still early access as of this writing

Gaps and Limitations

Max mode is early access. The most compelling capability — multi-step agentic planning across tools — isn’t GA yet. The current Rovo Chat is useful but not transformative.

Rovo Dev is a separate line item. The developer experience requires an additional $20/developer purchase. For a team using GitHub Copilot or Cursor, adding Rovo Dev requires justifying the overlap.

Credit opacity. The credit system (25/70/150 per user/month) is not well-documented in terms of which actions cost what. Atlassian’s 90-day overage notice is reassuring, but teams on Standard plans may hit limits without clear visibility into where.

Agent reliability at scale is unproven. The 7x growth in agentic automations is a strong signal, but most published data is from Atlassian’s internal 80,000-developer cohort. Third-party reliability data on Rovo Agents in production workflows is limited.

Rovo Studio complexity for custom agents. “No-code” applies to simple templates. More complex agent workflows with conditional logic and custom integrations will push teams toward developer-level work.

Builder Implications

If you’re building AI tools for enterprise teams:

Rovo is now your competition inside Atlassian. If your use case is “help teams manage Jira tickets” or “help teams write Confluence documentation,” Rovo does that today — bundled, no additional procurement. The window for specialized Atlassian tooling is narrowing.

Rovo is also a platform. The Rovo MCP Server and Teamwork Graph CLI let your agents read Atlassian context without replicating it. If you’re building an orchestration layer that needs to understand what teams are working on, you can query the Teamwork Graph rather than building your own Jira integration.

The FDE pattern is relevant here. Atlassian is positioning Rovo Studio as the tool that lets enterprise teams build their own agents — but complex deployments will still need human guidance. If you’re in AI implementation services, helping teams configure Rovo Studio correctly is a legitimate engagement scope, not a threat to your work.