Is Rasa Stories performs better than a rules based deterministic dialog manager

Rasa Open Source
1

I am developing a chatbot in insurance domain, where there are dynamic range of intents and lot of stories have been trained.

Can rasa stories helps to predict an undefiled path in the conversations? What if a story is not defined, will it be able to identify the correct path?

2

sorry, I don’t understand what do you mean

3

I am building one chatbot but in two different way:

  1. I will use only the rasa nlu model to predict the intent and implement the dialog manager (some rule based deterministic dialog flow) out side rasa (with out using stories and domain).

  2. OR I can develop the same bot entirely using nlu, stories, rules, custom action.

Which will perform better? is it #1 or #2 ? Is it ok to develop the bot without using stories and rule?

4

it is impossible to say which will perform better, it depends on how good you implement it. We recommend to use all rasa capabilities including stories, rules etc