Hello Everyone,
We are developing a financial assistant bot to assist our clients’ (major banks’) call centers in reducing work load by deploying chatbots to resolve common issues such as balance inquiry, profile update, and card activation/block. We have completed a proof of concept for a single client but want to scale it for multiple clients. The issue is that different clients’ conversation flows and stories differ slightly. One approach is to create a separate model for each client, which is not feasible in our scenario due to the large number of clients and the difficulty of managing their training data and models on servers. Can we create a generic model in Rasa that allows us to skip the conversation or slot filling in the case of form filling based on client id? Please share your thoughts.
*** example:**
if user wants to block the card one bank needs following information
-
card number -
social security number of card-holder -
card pin
other bank needs following information:
-
card number -
social security number of card-holder
and we want to deal the above scenario with single story for 2nd bank we don’t want the bot asks about the card pin in order to block the card