FAQ/Conversational hybrid bot

[AaronDeRybelHowest]

I’m currently trying to create a chatbot that can handle some predefined FAQ but also is able to hold some basic conversation using the ML based approach.

I want to:

1. Always give same response to a list of FAQ.
2. Also have different intents than “FAQ” that can be used in conversation.

I’ve followed the tutorial on FAQ bots: Chitchat and FAQs. It only allows to make a different responseselector for each intent = FAQ, chitchat.

Has anyone tried to make a hybrid bot like this ?

[jjuzl (Joe Juzl)]

Hi @AaronDeRybelHowest

Yes! There is nothing stopping you using a rule policy for things like an FAQ but training stories for more general conversation.

See: GitHub - RasaHQ/rasa-demo: Sara - the Rasa Demo Bot: An example of a contextual AI assistant built with the open source Rasa Stack

[AaronDeRybelHowest]

Hi @jjuzl ,

I’m determined to get to the bottom of how retrieval intents work ! My best guess now is that i’m missing custom actions.

Right now i’m trying to reproduce a minimal version of Chitchat and FAQs So I added some FAQ/chitchat (it’s for the dutch language).

1. When running rasa data validate I get errors:
   
   
2. When running rasa train I get errors for every single sub intent of faq and chitchat:
   

My best guess now is I need to add custom actions called utter_faq and utter_chitchat that map to the subintents. I’ve seen some custom actions on the example chatbot you linked. But I can’t figure it out in my example, how I would make this?

for reference these are my files
domain.yml
config.yml
rules.yml
stories.yml
nlu.yml