As per my knowledge, there are 2 type of Chatbots- Open Domain and Closed Doman Intent Based. I was wondering how are conversational agents created such that can converse as if they are Open Domain, but still respond to specific user queries and intervene. As an example, consider a healthcare bot that acts as an empathetic conversation agent but also tells you specifically when to take medicine and answers specific questions about the medicine.
The obvious approach that comes to mind is fine tune something like DialogGPT etc. for your own purpose and also have an intent classifier. If the intent classifier does not recognize an intent then pass control to DialogGPT.
The problem with this is once "inside" an intent you have to be careful in ensuring the conversation related to the intent is over before proceeding further. Additionally you would have to pass in the whole context to DialogGPT each time which would be fairly time consuming. I was wondering how chatbots and personal assistants overcome this problem. Any help would be appreciated!