Scribe analytics

The Scribe transcription solution is able to extract additional information from the transcribed text such as key terms, key phrases, quotes, and product classes. The information is stored in standard metadata fields and can be access at various playec in the system. It can be used during search, data management policy configuration, labeling rules, export configuration, etc.

Enabling analytics for Scribe

In order to use the analytics capabilities of Scribe, the related features have to be enabled in the data processor configuration, and Scribe Docker containers have to be deployed with the right configuration. Certain features might require additional licenses. For more information contact support. To enable analytics in the data processor configuration, follow the steps below: 

Step 1 - On the web interface go to Data / Data Processors.  

Step 2 - Select the already existing data processor or add a new one by clicking on the Add new Data Processor link.

Step 3 - Verify the Name for the data processor.

Step 4 - Verify the Type and select the Speech Transcription. 

Step 5 - Verify Engine and select Scribe.  

Step 6 - Choose the analytics features you want to use. Some features might require additional licenses. 

Enable Speaker DiarisationSeparates the speaker in the transcript
Enable Gender ClassificationNot supported
Enable Identify QuotesIdentify quotes by creating the IM text of the quotes and identify the product class traded
Enable InsightsExtracts key terms and key phrases


Step 7 -
Click Save to save the data processor.

Once the analytics features are enabled, the Verba system will call the Scribe APIs and ask for the additional analytics information. The Scribe Docker containers have to be configured accordingly in order to produce the analytics. 

Accessing analytics in search

Configuring the conversation list layout

In order to access the analytics in search, the search layout has to be changed as follows:

Step 1 - On the web interface go to Conversations.

Step 2 - Click on the icon in the top right corner to configure the conversation list layout. This will change the layout for the logged in user only. Administrators can also change the default list layout for all users (only before the users are created).

Step 3 - Select and add the customer metadata fields labeled as Scribe | ... 

Step 4 - Click on Save.

Step 5 - Optionally, the appearance of the custom metadata fields can be changed under System \ Metadata Templates. Administrators can define if the metadata field should appear as a column or as a new line in the search grid. The maximum length of the fields can also be defined.

Searching

Standard metadata search is available for Scribe related analytics:

Step 1 - On the web interface go to Conversations.

Step 2 - Drill down the Metadata and Markers section.

Step 3 - Under Metadata fields, select one of the fields labeled as Scribe | ... 

Step 4 - Select the operation and define the filter criteria.

Step 5 - Press Search to execute the query.

Analytics in player

The advanced analytics information is available in the player as well:

InformationDescription
Key termsDisplayed on the Transcription and on the Analytics tab. Clicking on the terms, the system highlights all occurrences of the term in the transcript.
Key phrasesDisplayed on the Analytics tab.
Quotes

Quotes are marked on the Transcription tab with a icon. On mouse over, the IM text of the quote and the identified product class is displayed.

On the Analytics tab, detailed information for all identified quotes is displayed. It shows the timestamp, the IM text of the quote, the product class, and the related original transcript. By clicking on the quote, the player jumps to the beginning of the quote in the recording.

Product ClassThe information is displayed on the Analytics tab under the Quotes section.

Transcription tab

The transcription tab shows the full transcript with speaker separated (when enabled). The speakers are indicated as Speaker 1, Speaker 2, etc.

Analytics tab

The analytics tab shows advanced analytics for the transcription.