Back to Changelog
New
AI Agent Supervise Mode
We've introduced a new Human-in-the-loop mode for AI Agents, called Supervise Mode, allowing teams to review and approve AI-suggested actions before they go live.
What's new
Enable Supervise mode at the channel level.
AI Agents will suggest replies, escalations, or label assignments.
These suggestions require human approval before execution.
All approved actions appear under the human agent's name.
Why this matters
You can now confidently test and refine AI Agent behavior in a live environment, without risking unintended customer interactions. Build trust in your AI's accuracy and performance before moving to full automation.
Like this update? Let the team know so they get even more motivated to bring more awesome updates to you.