Implementing a cognitive chatbot using CRLMM, as for any chatbot, needs to follow some well-organized steps:
- Identify the intents and entities
- Create the dialog scenarios
- Dialog controls—all that the user clicks and answers
- Make an emoticon images repository
- Make a conceptual images repository
- Train and test the model with all sorts of scenarios; especially look out for the ones that fail to improve the system