Reading Call Transcripts

The Conversations section in your dashboard includes a full text transcript for every AI-handled call. Transcripts let you review exactly what was said without needing to listen to the audio recording.

Where to find transcripts

  1. Go to Dashboard Home or AI Receptionists → [Your AI] → Conversations.
  2. Click any call to open the conversation detail panel.
  3. Click the Transcription tab.

What the transcript shows

The transcription tab displays the conversation as a series of messages:

  • Agent messages (what the AI said) — shown in one style.
  • User messages (what the caller said) — shown in another.

Each message includes a timestamp showing when in the call that message occurred (e.g. "1m 23s" means the message started 1 minute 23 seconds into the call).

How to use transcripts effectively

Quality check: Read transcripts for calls that were marked as "Failed" or where a booking wasn't completed. Often, the transcript reveals a specific point where the conversation broke down — a misunderstanding, an unclear question, or a gap in your knowledge base.

Knowledge base improvement: If the AI answered a question incorrectly, the transcript shows exactly what the caller asked and what the AI said. Use this to identify which knowledge articles need updating.

Greeting review: The transcript starts from the AI's first message. You can see exactly how your greeting sounds in the context of a real call — sometimes issues are only obvious when you read the full conversation.

Staff training: Share transcripts with staff to help them understand how the AI handles different call scenarios. This builds confidence in the AI's capabilities and helps staff know when the AI might need manual backup.

The call summary

Before the transcript, the Overview tab shows an AI-generated summary of the call. This is a helpful quick-read if you're reviewing many calls and want to triage which ones need closer attention.

💡 Tip: Set aside 10 minutes a week to read a few random call transcripts. You'll often discover something unexpected — a question customers ask that you hadn't anticipated, or an area where the AI could be improved.