Discussing the hybrid approach

In a real-life scenario, in order to build the chatbot we can also combine some of the techniques described here. As per the business needs we can use a hybrid approach.

Let's take an example. Suppose you are building a chatbot for the finance domain. If a user asks for the available balance in his account then we just need a rule-based system, which can query the database and generate the account balance details for that user. If a user asks how he can transfer money from one account to the other account, the chatbot can help the user by generating step-by-step information on how to transfer money. Here, we will use the deep learning-based generative approach. We should have one system that includes a rule-based engine as well as a deep learning algorithm to generate the best possible output. In this system, a user's question first goes to the rule-based system. If that question's answer can be generated by the rule-based system, then the answer will be passed...