Help Center

AI Call Participation Metrics

AI Participation Metrics analyze your recorded calls to show how the conversation actually flowed. Instead of guessing who talked the most or whether the discussion was balanced, you get clear, objective insights based on the call transcript.

This feature helps you understand:

• Talk time balance — Whether the conversation was a healthy back-and-forth
• Who spoke the most — Percentage breakdown for each participant
• Conversation dynamics — Interruptions, pauses, and turn-taking patterns
• Engagement signals — How actively each person participated

Why Participation Metrics Matter

Great conversations usually have good balance. If you’re talking most of the time, you may be missing important client insights. If the other person dominates the conversation, you may not be providing enough guidance.

Participation Metrics help you find the right balance and continuously improve how you communicate.

What You Need Before Generating Metrics

Before participation metrics can be generated:

• The call must be recorded
• The call must be transcribed

If transcription is missing, the AI will not be able to analyze participation.

How to Generate Participation Metrics

1. Go to Contacts.
2. Open the contact associated with the call.
3. Click the Calls tab.
4. Locate the call you want to analyze.
5. Click the three dots menu (⋮) next to the call.
6. Select Generate participation metrics.

The AI will begin analyzing the call immediately.

Where to Find Participation Metrics

Once analysis is complete (usually within a minute or two):

• You’ll receive an email notification when results are ready
• A participation metrics note is added to the contact’s record

The note includes:

• Talk time percentages for each participant
• Clear breakdowns such as “Advisor spoke 45%” and “Client spoke 55%”
• Additional insights into conversation flow and engagement

Troubleshooting

• “Call transcription not found”
  The call must be transcribed before participation metrics can be generated.
• Metrics don’t look accurate
  Very short calls or poor audio quality can reduce speaker detection accuracy.