I would like to know more details about building Pega Web Chatbot and Facebook Chatbot.
1. What rules should I use for analyzing text? Of course I know that text analyzer, but how to make this chatbot smart? Can I use for this Data Set, Data flows etc., or they should be used only for analyzing posts on Facebook, Twitter... . If I can use it, how I should build this and what are good practices for this?
2. When I want to run the case from chatbot what is a good practise of doing that, I should create new one case type, only for chatbot, or I should add Facebook or Web Chatbot channel to stages in my existing cases(have separate parallel process which is run based on "IsWebchat" condition)?
3. Finally, I should create separate case for all chatbot conversation or it's not necessary?
1) Starting 8.3, you can use Dataflows, strategies, etc to provide more contextual responses in your chatbot. You will need to switch to iAPI based text analysis technique (default for new channels) in channel configuration and you can override pyInteractionDF and pyGetBestPropositions rules in corresponding channel class. Note, each channel has its own Pega Class under Data-Decision-Request where channel specific overrides can be done.
2) Good practice is to add parallel process for each channel so you can have channel specific questions.
3) Can you elaborate on this question? What do you mean by 'create separate case for all chatbot conversation'?