AI
Turning an AI agent on or off
Including how to put it into read-only mode while you review transcripts.
- AI
- 3 min read
Every AI agent we build has a single switch, and turning it off never deletes anything — conversation history, training and configuration all stay exactly as they were for whenever you turn it back on.
Where the switch is
From your dashboard, AI assistant → [agent name] → Status. Toggling it off stops the agent from replying to new messages immediately; anything already in a conversation stays visible, and a human on your team can pick it up manually from that point.
Read-only mode
Between fully on and fully off, there is a middle setting: the agent keeps reading incoming messages and drafts a suggested reply, but nothing sends until a person on your team approves it. This is what we recommend for the first week or two after any change to how an agent is trained — it lets you review real transcripts before trusting it to reply unsupervised again. See reading agent transcripts and escalations for what that review actually looks like.
Support
Talk to us directly.
Any of the three reaches the same team.
Ready when you are.
Tell us what you're building and we'll map the shortest route there.