Question about participant timestamps and mapping to call recording

This is a question about identifying the points in an audio recording that map to respective participant segments in a call.

Using this hypothetical scenario:

  • conversationStart field in conversation details (GET /conversations/{conversationId}/details) is 3PM today - 2023-10-09T15:00:00
  • Call details (GET /conversations/calls/{conversationId}) reveals two participants: 30 seconds of IVR (start: 2023-10-09T15:00:00, end: 2023-10-09T15:00:30) and two minutes of an agent interacting (start 2023-10-09T15:00:31, end: 2023-10-09T15:02:30)
  • IVR is not part of the recorded call audio.

The problem we face is we are unable to identify any data that expresses whether or not the IVR is part of a call's recording. Is there something that can inform us of this at run time, or is it something that must be known and planned for ahead of time?

Follow-on scenario: the same would apply for other non-human participants such as ACD.

The recording's metadata has the recording start time and end time. If you didn't record the IVR segment of the call, then the recording start time and the conversation start time would be different.

I stand corrected. Thanks @Jerome.Saint-Marc. The recording actually "starts" at a paused state for the IVR segment, so the recording timestamp would've been the same as the conversation timestamp.
The way to identify would be to look for annotations in the recording. If there was IVR recording suppression, there would be an annotation that explains the start, duration and reason of the recording pause.

This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.