Atlassian has turned on default data collection across its suite to feed its AI features. This shift means every user interaction—issues, comments, code snippets—will be automatically harvested unless opted out. By centralizing raw usage signals, the company hopes to accelerate model refinement for products like Jira and Confluence, where contextual understanding drives productivity gains. This policy activates on all cloud tenants.

The move follows a broader industry push to leverage internal telemetry for generative assistance. Atlassian argues that richer training data reduces hallucination and improves relevance when suggesting tickets, automating status updates, or drafting documentation. Administrators retain control through a new opt‑out toggle in admin settings, and the company promises compliance with existing privacy policies and regional data‑handling regulations today.

Critics warn that default collection could expose sensitive project details, especially in regulated sectors. Atlassian counters that all harvested content is anonymized and stored in encrypted buckets before being fed to its large‑language‑model pipelines. Early internal benchmarks suggest a measurable lift in suggestion accuracy, giving teams faster turn‑around on routine tasks while keeping the data pipeline transparent to end users.