Can we build a smart reply suggestion with rasa using a bunch of conversations? if yes, then how ?
Hi @shashank34. I am not 100% sure I fully understand your question. You should be able to build an assistant using conversational data. All you have to do with it is defining intents with some example inputs to train the NLU model. In addition to that you should create some stories data based on the conversations you have to build train the dialogue model.
We have a step-by-step tutorial on this process on our docs. Did you have a chance to look into it?
@Juste I had a look on this process. but in this we have to add nlu examples and conversations manually and that is very hard for any developer. I am looking, for example we have a chat conversation record of two persons then using that conversation rasa bot trained and then giving smart replies suggestions on any new conversation like a linkedin doing.