After deploying the servers and enabling the Speech Analytics Service, the system can be configured to transcribe voice calls. The configuration consists of the following steps:

• Creating a speech transcription data processor which enables the specific transcription integration in the system. The system supports multiple transcription engines. Multiple data processors can be created and enabled. 
• Assigning transcription licenses to users to allow transcribing the calls of the users. Different transcription integrations may require different licenses.
• Creating speech transcription data management policies to define which calls will be transcribed by which processor and which ASR model.

Creating a Data Processor

A data processor represents the connection with the transcription engine. In order to set up a processor, follow the steps below. 

Step 1 - In the Verba Web Interface go to Data / Data Processors  

Step 2 - In the top right corner click on the Add new Data Processor link

Step 3 - Enter the name of the data processor 

Step 4 - Select Speech Transcription for Type 

Step 5 - Select the Engine 

Step 6 - Follow the guidelines specific to the transcription integration to configure the connection parameters:

• Verint transcription
  
• Intelligent Voice transcription
  
• Speechmatics transcription
  

Step 7 - Click Save to save your data processor 

Assigning transcription user licenses

Transcription is licensed on a per-user basis. You can assign these licenses to roles (and the roles to users) by enabling one of the Speech Analytics / Transcribe Conversations checkboxes on the role configuration page. The transcription policy only selects calls that belong to users where there is transaction permission enabled, otherwise the call will not be transcribed. Certain integrations have specific license requirements or additional limitations. For more information on these, refer to the integration specific page.

Creating a Transcription Policy

A Data Management policy configured with the Speech Transcription action is used for specifying the recorded conversations to be transcribed. Servers running the Verba Speech Analytics Service read the media file from the storage, then store the recorded conversations on their local disk temporarily while transcribing. After the transcript is created, the media files will be discarded, and the transcript files are relocated to the media's storage location. Each policy you define can be set up for different ASR models and a set of filters to determine which conversations should be transcribed. These filters are based on the metadata stored in the database for each conversation.

This is a good way to account for differences in languages used for conversations of different locations, groups, etc.

Step 1 - Follow the generic policy creation steps described on the following page: Data management policies

Step 2 - Select Speech Transcription as the action

Step 3 - Select the Data Processor

Step 4 - Select the Language for the transcription

Step 5 - Select your filters to specify which conversations should be checked by this policy

Step 6 - Click on Save. The policy will run periodically,

Running and monitoring the transcription process

The transcription actions are executed by the Verba Speech Analytics Service on the servers where this service is enabled. 

The transcription process can be monitored on the Dashboard with a Background Tasks Widget or in the list of background tasks in the Verba Web Interface at System / Background Tasks.

The Log files for the service can be found at (your application path might be different):