Atlassian began rolling out new “data contribution settings” across its admin console on April 16, documenting a policy change that will allow the company to use customer data for AI model training starting August 17, 2026. The settings, detailed in official Atlassian support documentation, cover two categories of data: metadata (issue types, project structures, usage patterns) and in-app content (comments, descriptions, and content created in Jira, Confluence, and other Atlassian tools).
Who Is Opted In by Default
The defaults depend on your plan tier, according to Atlassian’s documentation:
Free and Standard plans: In-app data contribution is on by default. Admins can opt out. Metadata is always contributed and cannot be turned off.
Premium plans: In-app data contribution is off by default. Admins can opt in. Metadata is always contributed and cannot be turned off.
Enterprise plans: In-app data contribution is off by default. Admins can opt in. Metadata contribution is on by default but can be turned off. Enterprise is the only tier with the ability to opt out of metadata collection.
The data will be used to train Rovo and Rovo Dev, Atlassian’s AI products, according to Agent Wars. Atlassian states on its Trust Center FAQ that all contributed data is de-identified and aggregated before use.
The Policy Reversal
This represents a reversal of Atlassian’s prior stated position. As recently as November 2025, Atlassian’s official support documentation stated that “customer data, including data submitted to or generated by Rovo and Atlassian Intelligence, is never used to train, fine-tune, or improve AI models or services.” The new documentation replaces that commitment with a tiered opt-in/opt-out framework.
The April 16 rollout gives organizations a four-month window to review and adjust settings before Atlassian begins using data on August 17. The company is hosting a webinar on April 28, 2026 to walk customers through the changes, per Agent Wars.
The Scope
Jira and Confluence are foundational tools for software development teams worldwide. Jira tracks bugs, feature requests, sprints, and deployments. Confluence hosts internal documentation, architecture decisions, post-mortems, and company knowledge bases. For teams building AI agents that integrate with Jira or Confluence (a common pattern for engineering workflow automation), the data flowing through those tools now has a dual purpose: operational use and AI training input.
The change follows similar moves by Microsoft (Copilot/M365 data policies), Salesforce (Einstein data use), and Google. Atlassian notes this pattern in its own FAQ.
For engineering teams on Free or Standard plans: check your organization’s data contribution settings at admin.atlassian.com before August 17, or the defaults apply.