Atlassian's AI layer turns on by default once a plan qualifies

The setup

Atlassian folded its generative features, originally branded Atlassian Intelligence, into a wider AI layer the company now calls Rovo, spanning Jira, Confluence, Jira Service Management, and Jira Product Discovery. The vendor's own admin documentation, as it reads on 16 September 2026, states that the features are "available and automatically enabled for all apps on Standard, Premium, and Enterprise plans." Reaching that state does not require a separate purchase decision at the point of rollout; it requires already being on one of those three plan tiers, with the Free tier and Government organizations excluded.

What the documents show

The admin guide describes a single settings path, "Rovo, then Rovo access in Atlassian Administration," through which an organization administrator manages preferences rather than performs an initial activation, since the default state is already on. A companion trust-center page separates two data claims: the outside large-language-model providers Atlassian uses "do not retain your inputs and outputs, or use them to improve their services," while Atlassian itself "may use metadata to fine-tune open-source models that operate strictly within Atlassian's infrastructure" for its own quality purposes. Both are vendor claims about vendor-controlled infrastructure; neither page cites independent verification.

The friction

The friction here is opt-out rather than opt-in. Because the feature ships on by default for qualifying plans, an administrator who wants to keep it off has to find the Rovo access setting before end users encounter it, not before rollout. The trust page's second data claim is conditioned on "data contribution settings," language that implies a separate, adjustable control the two pages do not spell out together. Sandbox environments need a distinct manual enablement step, and Government organizations cannot access the feature at all, a hard product boundary rather than an administrator's choice.

What changed in the work

For a team already paying for Standard or higher, the setup cost moved from a rollout project to a review task: instead of scheduling an enablement, an admin now audits a default that is already live. That is an editorial read of the documents' combined effect, not a claim either page makes directly. The rebrand from "Atlassian Intelligence" to "Rovo" as the umbrella name is also worth flagging for anyone searching the original product name and finding it absorbed into newer marketing.

• Is your organization's plan tier one where the feature is already on?
• Where does your admin console show the data-contribution setting the trust page references?
• Does your sandbox or Government-tier environment need the separate manual step the documentation describes?

The setup story here is less about a launch event than about a default: once a plan qualifies, the feature is already running, and the administrative work is deciding whether to leave it that way.