Skip to main contentDeployment is when you allow the AI to send replies on its own.
Users control both the level and rules for automation implementation.
Automation Levels
The system offers three deployment modes:
- Off - AI doesn’t propose or send replies
- Copilot - AI drafts, you approve or edit
- Autopilot - AI sends confident replies automatically, escalates the rest
Safety Controls
Implementation includes protective measures:
- Delays - Hold automatic sends for review
- Working hours - Restrict automation to certain times
- Urgent tags - Force certain types of messages to escalate, regardless of confidence
By repeating the Teach > Test > Deploy cycle, you can steadily expand what the AI handles, moving closer to full automation without losing oversight.