Behaviors & coaching
Behaviors that book meetings
When calls have transcripts (from an AI agent, or pasted from Gong, Zoom or a "Name: text" export), Clarity measures:
- Talk share: how much of the call the rep spoke
- Questions asked, monologues over 30 seconds, filler words
- Objections raised
- Eight call-flow checks, for example: asked permission, gave a reason for calling, offered a specific time
Only the measurements are kept, never the words.
On your home page, Behaviors that book meetings compares the meeting rate of conversations with and without each behavior. It only counts conversations (connected calls and meetings). Including unanswered dials would make anything that happens on a call look like it books meetings.
A behavior is only called a winner at 95% confidence. Until then it's marked still learning, with the rates side by side.
Coach this week
Admins see Coach this week on the Admin Dashboard: a short list of reps to talk to and why. It flags things like:
- Connects, but doesn't book: good conversations that don't turn into meetings
- Not reaching people: a connect rate well below the team's
- Losing to one objection: the same pushback keeps ending their calls
- Rarely offers a specific time when asking for the meeting
- Dials are down this week
It also calls out wins, such as a rep booking well above the team, so you know who to learn from. Each item shows its evidence and a suggested action, and only appears once there's enough data to be fair to the rep. It covers human reps, not AI agents.
Best time to call
When your team connects shows the connect rate by hour, in your local time, so you can plan call blocks around the hours that actually reach people.
