I have been recently exploring Rasa and how I can use it to build a banking chatbot. I have a fair understanding of how it works in general.
However, I want to go deeper and build a sophisticated bot that can answer any general questions about a bank, such as what are the latest card offers?, what is the interest rate for housing loan?, what is the forex rate?, what documents do I need to apply for a loan? … and so on. Likewise there can be any question that a customer can ask.
My question is :
- Is it possible to connect to a bank’s realtime database and fetch the necessary information based on user’s query? For eg. if a user asks what are the latest card offers, the bot should be able to query the database and look for card_offers table and fetch relevant data. How to prepare the training set for this scenario ? And how to modify stories, nlu and domain file ?