Enhance or create a voiceprint model using a recorded interaction

Search the system for interactions to create a voiceprint model or to provide more training to improve the model.

Before you begin

Create a person

Procedure

Use the search feature of Risk Management to locate a recorded interaction you want to use for voiceprint model training.
In the search results list, select the interaction you want to use for the voiceprint model.
Click add selected contact to voiceprint model .

The Add to Voice Model dialog box is displayed. The Add to options are based on the attributes associated with the interaction.

Select the correct channel. If you choose the wrong channel, the integrity of the voiceprint model is compromised. For help with determining which channel to select, use Playback and adjust the balance between channels.
Select which voiceprint model to update or create with the Add to option. Use one of the following:
- Employee: if the system can identify an employee within the Risk Management system that is on the interaction, it populates with the employee name.
- Customer: if the system can identify a customer within the Risk Management system that is on the interaction, it populates with the customer name.
- Target: if the system can identify a target within the Risk Management system that is on the interaction, it populates with the target name.
  
  If the system cannot identify speakers on the interaction, then [Not Available] displays. Use the Select option to associate a person.
  
  If a person is already selected, but you need to change the person, click the Change option to select a different person.
For a stereo recording, select the Channel containing the voice you want to use in the model. Select from:

Stereo Left / Mono: Use the stereo left / mono channel for what is typically the employee voice in stereo recording environments. Always use the left channel for mono recording environments.
Stereo Right: Use the stereo right channel for what is typically the customer voice in stereo recording environments.

Users with the privilege to Approve/Unapprove Interactions for Enrollment can select or clear Approve Interaction. The option is selected by default. If selected, the interaction is automatically approved for use by the voiceprint model.

If you do not have the privilege to approve audio for voiceprint models, the approval option does not display. Manual approval is required.

If not added in an approved state, audio is added to the person in a Pending Approval state.

A user with the privilege to approve and unapprove interactions must manually review and approve the interaction before it is included in the voiceprint model.
After you have selected the voiceprint model to update, click OK.