After you create a Pega Intelligent Virtual Assistant (IVA), you can train the natural language processing (NLP) model with sample data to improve how the IVA responds to user interactions. The NLP model for the configured conversational channel is saved as a text analyzer rule. An IVA uses this rule to analyze any text of the chat conversation for sentiment analysis, text subject (topic) classification, intent analysis, detected language, and entity extraction.

For more information, see Updating the NLP model for a Facebook channel, Facebook channel NLP model, Updating the NLP model for a Web Chatbot channel, and Web Chatbot channel NLP model.