Adding a text analyzer for an IVA
To enhance the artificial intelligence of Pega Intelligent Virtual Assistant™ (IVA) so that the system provides better responses and detects entities and the topic, the subject matter, you can configure one or more advanced text analyzers together for an application. In that case, the text analyzers examine user input one by one, until the system finds a response and detects the correct data.
You can also configure one text analyzer to run only within the case context, and another one to run outside of the case context. For example, to refine interaction with a user when a case is started in the IVA, you can specify the iNLP advanced text analyzer to improve the text analysis of user input, while using the simple exact match text analyzer if the system does not start a case during a chat session.- In the navigation pane of App Studio, click Channels.
- In the Current channel interfaces section, click the icon for your existing Digital Messaging or Alexa channel.
- In the channel, click the Behavior tab.
- In the Text Analyzer section, select the Use advanced configuration checkbox.
- In the Text Analyzer section, select a method for
configuring a text analyzer:
- To create a text analyzer, click Add text analyzer.
- To edit an existing text analyzer, click the Switch to edit mode icon next to the text analyzer that you want to edit.
- In the Text Analyzer type list, select and configure a
text analyzer:
Choices Actions Exact match Configure the default text analyzer that exactly matches user input to a response. Select whether to use the text analyzer within the case context, outside of the case context, or both.
Pega NLP Configure an advanced text analyzer that uses the best approximate match by using advanced NLP and artificial intelligence: - Select or define a text analyzer rule for this definition type with the sentiment, classification, topic, and entity extraction analysis.
- Select whether you want to detect entities or topics within the case context, outside of the case context, or both.
- To ensure that the chatbot engages users in casual conversations, select the Enable small talk checkbox.
- In the Text analyzer rule field, create or select a text analyzer rule.
- To modify the text analyzer rule settings, click the
Open Text Prediction
field.
For more information, see Building text analyzers.
- To modify the text analyzer rule settings, click
Open text analyzer in Dev Studio, make changes,
and then in the lower-left of the page click Back
to Pega App Studio.
The option to modify the text analyzer rule is only available if you have access to Dev Studio.
For more information, see Building text analyzers.
iNLP Configure an advanced intelligent NLP text analyzer / text prediction that uses adaptive analytics text analysis: - Select whether you want to detect entities or topics within the case context, outside of the case context, or both.
- To allow users to add training data for the text models, select the Enable model training checkbox.
- To ensure that the chatbot engages users in casual conversations, select the Enable small talk checkbox.
- To modify the text prediction settings, click
Open Text Prediction.For more information, see Analyzing messages with text predictions.
- To modify the text analyzer rule that underlies the text
prediction, click Open Text Analyzer in Dev Studio, make changes,
and then in the lower-left of the page, click
Back to Pega App Studio.For more information, see Building text analyzers.
This type of analysis integrates text analytics with strategies, propositions, and interaction history to provide the context for making better next-best-action decisions. For more information, see Customizable Interaction API for text analytics.
- Click Submit.
For example: The following figure shows the advanced configuration section for the iNLP text analyzer for a chatbot: The advanced text analyzer configuration section for an IVA 
- Optional: To add or configure more text analyzers for the IVA, repeat steps 5 through 7.
- Click Save.
Previous topic Configuring a text prediction for an IVA Next topic Defining topics for text analysis for an IVA