Settings

search⌘K

Settings section

Workspace and AI rules

Configure workspace identity, regional timezone, and autonomous AI reply guardrails.

How this works:AI responses below your confidence threshold or matching always-ask topics are held for human review in the attention queue.

Workspace name

Timezone (persists in UTC)

Draft-first mode

Autonomous confidence threshold (85%)High accuracy

Replies generated with confidence below this threshold are placed in the human attention queue for one-click approval.

Enable autonomous auto-send for high-confidence responses

When turned on, responses that exceed your confidence threshold and clear policy checks are dispatched directly.

Always-ask-a-person topics

Inquiries regarding these sensitive topics always require human agent verification regardless of AI confidence.

refunds and billing disputescontract legal termssecurity vulnerabilitiesaccount cancellations