What Is Atlassian Rovo? (And How It Differs from Atlassian Intelligence)
Atlassian Rovo is the cross-platform AI layer for Atlassian Cloud. It does three things: Search across Jira, Confluence, and connected SaaS tools; Chat that answers questions grounded in your organization's knowledge; and Agents that carry out multi-step tasks autonomously. It is different from Atlassian Intelligence, which is the in-product AI engine — Rovo uses that engine and extends it across your entire toolchain.
- What Rovo actually does (Search, Chat, Agents, Studio)
- The precise difference between Rovo and Atlassian Intelligence
- Which data sources Rovo can reach
- How Rovo Agents work and what they can do
- What Rovo means for admins and how to prepare your content for it
If you have heard the name "Atlassian Rovo" and are not sure whether it is a new product, a rebrand of Atlassian Intelligence, or something your team should adopt, you are not alone. The naming overlap confuses most teams. This guide explains what Rovo is, how it relates to the AI features you may already be using, and what it changes for the people who administer Atlassian Cloud.
Rovo in one paragraph
Rovo is Atlassian's cross-platform generative AI product. Instead of living inside a single app, it sits above Jira, Confluence, and Jira Service Management — and above connected external tools like Google Drive, SharePoint, and GitHub — and lets you search, ask questions, and automate work across all of them at once. Think of it as the AI layer that finally makes your sprawling knowledge findable and actionable, rather than another feature buried inside one product.
Rovo is built on three pillars:
- Rovo Search — semantic search across Atlassian content and connected SaaS sources. You ask a question in natural language and get answers with citations pointing to the exact page, issue, or document the answer came from.
- Rovo Chat — a conversational interface that synthesizes answers from across your tools, so you can ask "what is the status of the Q3 migration?" and get a grounded answer pulling from Jira epics, Confluence decisions, and a linked Slack thread.
- Rovo Agents — autonomous AI teammates that carry out multi-step tasks: triaging incoming issues, drafting release notes from completed tickets, compiling a weekly project status report from scattered sources.
A fourth surface, Rovo Studio, is Atlassian's builder tool for composing custom agents, automations, and apps on top of the Rovo platform.
Rovo vs Atlassian Intelligence: the key distinction
This is the most common point of confusion, so it is worth getting exactly right.
Atlassian Intelligence is the AI engine embedded inside individual products. It is the feature that summarizes a Confluence page when you click a button, drafts a Jira comment from a prompt, or turns a rough issue description into structured acceptance criteria. It lives inside each product and acts on the content of that product.
Rovo is the cross-platform product that sits above all of them. It uses Atlassian Intelligence as its underlying engine, but extends it in three ways: it searches across multiple products and external SaaS tools, it lets you chat about combined knowledge from all of them, and it runs agents that act across product boundaries.
The simplest mental model: Intelligence is the brain inside each app; Rovo is the colleague who has read everything across all your apps and can act on it.
| Aspect | Atlassian Intelligence | Atlassian Rovo |
|---|---|---|
| Where it lives | Inside each product (Jira, Confluence, JSM) | A standalone layer above all products |
| Scope | Single product at a time | Cross-product + external SaaS tools |
| Core capability | In-product AI (summarize, draft, transform) | Search, chat, and autonomous agents |
| Data reach | The content of the product you're in | Atlassian + Google Drive, SharePoint, GitHub, and more |
| Relationship | The underlying AI engine | Uses Intelligence and extends it across tools |
If your team only ever works inside one product and only needs in-the-moment AI help, Intelligence may be enough. If your knowledge is spread across Jira, Confluence, and half a dozen SaaS tools and nobody can find anything, that is the problem Rovo solves.
What Rovo can search
This is where Rovo's value becomes concrete. Rovo Search reaches:
- Atlassian products — Jira issues, Confluence pages and spaces, Jira Service Management tickets and knowledge bases.
- External SaaS sources — Google Drive, Gmail, Google Calendar, SharePoint, OneDrive, Microsoft Teams, Outlook, Azure DevOps, GitHub, GitLab, Box, Dropbox, Figma, Loom, Lucid, Miro, Smartsheet, DocuSign, Zendesk, and Compass.
The practical impact is significant. A question like "where did we document the authentication redesign?" no longer requires knowing whether the answer is in a Confluence page, a Jira epic, a Google Doc, or a Figma file — Rovo searches all of them and returns a grounded answer with citations to the source.
For admins, this means content quality now matters across tools, not just inside Confluence. A well-titled Jira issue and a clearly named Google Doc are both discoverable by Rovo; a vague one is invisible to semantic search no matter how powerful the model.
How Rovo Agents work
Rovo Agents are the most consequential part of the product, because they move AI from "answers questions" to "does work."
An agent is given a goal and a set of trusted data sources, then executes the steps needed to achieve that goal — reading from the sources it was given, reasoning about what it found, and producing an output or taking an action. Crucially, agents are grounded: they work from your organization's actual knowledge rather than generating plausible-sounding guesses. That grounding is what makes them safe enough to delegate real work to.
Typical use cases teams are running today:
- Issue triage — an agent reads each incoming bug, categorizes it, suggests a priority, and assigns it based on the team's historical patterns.
- Status reporting — an agent compiles a weekly project update by reading completed issues, open blockers, and related Confluence decisions, then drafts the report for a human to review.
- Knowledge synthesis — an agent researches a topic across Jira, Confluence, and Google Drive and produces a briefing document with citations.
- Onboarding — an agent answers new-hire questions by searching across policies, team pages, and historical decisions.
Agents do not replace humans; they do the gathering and drafting that humans find tedious, leaving the judgment calls to people. The reduction in manual project work is real and measurable for teams that deploy them well.
What Rovo means for admins
If you administer Atlassian Cloud, Rovo changes three things about how you should think about your content and apps.
1. Content discoverability now depends on semantics, not navigation. Historically, a page was findable if it was linked from the right place or titled with the exact keyword. With Rovo, a page is findable if its meaning matches the question being asked. That rewards clear, substantive writing and penalizes vague titles and thin content. The cleanup work you may have been putting off — fixing bad page titles, merging duplicate pages, writing real descriptions — now has direct ROI because Rovo surfaces well-written content and buries weak content.
2. Permissions are respected, which means permission hygiene matters more. Rovo only returns answers a user is allowed to see — it inherits Atlassian's permission model. That is the right behavior, but it means broken or overly broad permissions produce wrong results: if a page is accidentally restricted, Rovo will not find it for people who need it; if a space is too open, Rovo may surface confidential content in answers. Auditing permissions becomes part of keeping AI useful.
3. Forge apps are more Rovo-friendly than Connect apps. Rovo indexes content that lives as native Atlassian data. Forge apps store their data on the Atlassian platform — in Confluence storage format, Jira entities, or Atlassian-managed storage — so their content is discoverable to Rovo. Connect apps often store content on vendor-hosted servers outside Atlassian, where Rovo cannot reach it. If AI discoverability of your app content matters, this is one more reason to prefer Forge. (For a deeper treatment of the Forge vs Connect split, see our Forge vs Connect guide.)
Licensing and availability
Rovo is included on Standard, Premium, and Enterprise Cloud plans for Jira, Confluence, Jira Service Management, and Teamwork Collections. AI is activated automatically on these plans — there is no separate Rovo license to procure for most Cloud customers. Some eligible plans are still in the process of gaining access as the rollout completes; if you do not see Rovo yet, it is likely on its way.
For Data Center customers, Rovo Data Center connectors for Jira, Confluence, and JSM enable data synchronization so cloud AI features can work with on-premises content.
A practical note for NGPILOT users
Every NGPILOT app is built on Forge, which means the content your team creates with them — diagrams rendered into Confluence pages, Markdown saved as native storage format, code blocks, FAQs, and structured content — lives as Atlassian data, not in an external vendor store. That is exactly the kind of content Rovo can index and ground its answers in. If you use NGPILOT apps to document your processes and systems, Rovo makes that documentation more valuable because it becomes findable and citable across your whole organization. Browse the app catalog or explore solutions by scenario to see how the apps fit the way your team works.
Frequently Asked Questions
Is Rovo the same as ChatGPT for Atlassian?
No. ChatGPT is a general-purpose chatbot with no access to your organization's data unless you explicitly connect it. Rovo is grounded in your Atlassian and connected SaaS content — every answer cites the source it came from, and permissions are respected so users only see what they are allowed to see. Rovo also runs agents that take action, not just answer questions.
Do I need to pay extra for Rovo?
For most Cloud customers, no. Rovo is included on Standard, Premium, and Enterprise Cloud plans for Jira, Confluence, JSM, and Teamwork Collections, with AI activated automatically. If you are on a lower plan or Data Center, check your specific entitlement.
Can Rovo replace our internal search tool?
For many teams, yes. Rovo Search replaces keyword-matching search with semantic search that understands meaning and returns grounded answers with citations. It reaches both Atlassian content and external SaaS sources, which most internal search tools cannot. The main caveat is permissions: Rovo is only as accurate as your permission model is clean.
Will Rovo agents make changes to my Jira or Confluence automatically?
Agents can take actions like assigning issues or drafting content, but you control what they are allowed to do. Most teams start with read-and-draft agents (research, summarize, compile) and gradually grant write actions as trust builds. The grounding in your actual data is what keeps agents from going off the rails.
Where can I read Atlassian's official Rovo documentation?
Atlassian's Rovo product page covers the core capabilities, and the Atlassian AI and Rovo community tracks real-world usage and updates.
NGPILOT builds exclusively on Forge, so content created with NGPILOT apps is native Atlassian data — discoverable by Rovo. Browse the app catalog or learn about how NGPILOT handles security and data.