Utterances are the content of the conversation. They describe who said what, and when. Each utterance must be attributed to an individual, so those individuals must exist prior to creating the utterances.

🚧

Deprecated fields

The deprecated "intervals", "interval_time", and "contexts" are related to "v1" analyses. See the "v2" Create Analysis API.

📘

Utterance Source IDs

Utterance source IDs are handled differently than source IDs for other entities. For utterances, source IDs are optional, but if provided, each source ID must be unique within the conversation.

📘

Ending a Meeting

If you include the time_end parameter in the body, the meeting will be ended once utterances have been collected. See End Conversation for validation on this value.

📘

Kicking off Analysis

The intervals and interval_time parameters are both optional. However, if included, only ONE of them may be specified. If neither are specified, (and the conversation is ended) the analysis job will determine a reasonable number of intervals for the analysis based on the length of the conversation.

📘

Validation

  • The time_start value must be equal-to-or-greater-than the time_start value of the conversation that the utterance belongs to.
  • The time_start value must be less-than the time_end value of the conversation that the utterance belongs to.
  • The time_end value must be equal-to-or-less-than the time_end value of the conversation that the utterance belongs to.
  • Status 406: Note that the maximum request size of 100KB is enforced.

📘

Analysis Contexts

You may provide one or more contexts for the analysis. The default context is just "basics", which includes:

  • about
  • follow_up
  • intent

If your deployment includes other contexts, specify those in the request as well.

This context will be used for the analysis of each individual, independently, in the conversation.

Language
Authentication
Header