Skip to main content
Your AI agent is live. Now you need to see what it’s actually doing. Every call your Voice AI agent takes, and every conversation your AI text agents handle, lands on the Conversations page — a real-time log of all customer interactions across your organization. That page is where you review your agent’s work, spot what it’s getting wrong, and tell us about it. Conversations

​
A first review pass

For a newly launched Voice AI agent, this is a good half hour:
1

Open the Voice tab and filter to your agent

Set the Agent Type filter to Chatbot, pick your agent, and set a date range covering since it went live. See Using the filters.
2

Start with the calls that should have booked

Filter Booking Status to Not booked — bookable calls where no booking was made. These are your missed opportunities, and the fastest route to improving the agent.
3

Check what it handed off

Filter Handling to Transferred to see which calls the agent passed to a person, and decide whether it should have handled them itself. See Live Transfers.
4

Listen to a few in full

Click into a conversation to play the recording, read the transcript, and check the Context tab. See How to Review Conversations.
5

Flag anything that went wrong

If a call didn’t go the way it should, flag it from the side drawer with a note. Your Hatch team picks it up and works it through to resolution. See Flagging a Call.
Reviewing calls regularly in the first few weeks is the single biggest thing you can do to improve a new agent. The metrics tell you which calls to listen to; flagging tells us what to fix.