Remove a voiceprint model audio segment

To control the integrity of audio files used to create a voiceprint model, you can remove a segment if it is an unacceptable audio source. A removed segment is permanently excluded from the model, which is retrained using remaining approved audio.

Before you begin

Create a person.
Enroll the person by creating a voiceprint model. You can use manual enrollment (by searching existing interactions or by uploading audio from an external source) or automatic enrollment (by a Recorder Analytics Rule).

Procedure

Access the Voiceprint Model for the person whose model you need to update.
In the Audio Segments list, select the audio segment you want to remove.
To remove an audio segment from the user’s profile, click remove .

A confirmation is displayed to verify the removal.
Click Yes.

When you remove an audio segment for an already trained model, the Training Status changes to Invalidated. The system will retrain the model using only approved audio segments. Retraining requires approximately 30-60 seconds as the audio is processed and the voiceprint model is trained.