What data AI features use, when they use it, and the limits around that processing.

This resource explains what context may be shared when AI-powered features run, what may be logged, and what controls users can expect.

Processed data types

User-triggered prompts, selected document snippets, tasks, or limited workspace context.

Request metadata and constrained telemetry used for reliability, abuse prevention, and debugging.

Model responses that may be logged temporarily for quality, audit, or product-improvement purposes.

AI calls are made only when the related feature is triggered; the product does not continuously scan private content in the background.

Sensitive or excessive context should be minimized, and downstream retention or deletion should stay aligned with the privacy policy.

AI features may send text or task context to third-party model providers and infrastructure processors.

The active AI processor list, data-region posture, and retention behavior should stay consistent with this page and the privacy policy.

Questions about AI-assisted processing can be routed through the privacy or support contact paths.

Legal resources

Privacy, AI, refund, permissions, and account deletion resources are public.

Notra