Hello to all,
I’m noob in RASA. I would like to know if RASA NLU would be suitable for my use case. I need to extract intentions and entities from a transcribed conversation between a customer service and its customers. We are not in the case of a chat bot but we do have a conversation. Example: customer service : hello what can I do for you customer: I would like an address of an Italian restaurant customer service : ok where are you ? customer : I am in New York etc…
So I have the whole conversation.
What makes me hesitate is the fact that from what I saw in Rasa NLU, we don’t have the context in the sense that it’s only the words of each interlocutor independently that are analyzed. The customer service note asks for codes that have the same format, so the only way to know what code it is and therefore the entity attached and that RASA NLU can take into account the request made by the customer service. example: Customer service: What is your customer code? customer: 124 566 33 Customer service: What is your special code? customer : 321 566 99
So, should we use examples during training with the last sentence of the customer service + the customer’s answer? Is it necessary to do otherwise, or is RASA NLU not necessarily the most appropriate solution and in this case which open source solution would be it?
Thanks in advance for your help.