There are times when your chat bots will struggle to narrow down the underlying user intent being expressed.
You need a way to find more signals for your chatbot to work with and to make it easier to pair a user query against a particular intent. This process is known as disambiguation.
For example, consider the following utterance: I'd like to apply an update to my bank account?
The utterance is very open ended and can mean many things to many people.
A human would need to ask some clarifying questions and you conversational AI solution is no different. The person might want to:
You have a few different ways to disambiguate user intent to help improve the accuracy of your conversational AI logic.
One of the first things you can do is to ask more clarifying questions. A follow up question from the earlier example might be: what type of update do you want to apply?
Set a limit on the number of clarifying questions you offer the person interacting with your conversational AI application as the person may have difficulty articulating their request. Too many questions will likely annoy users.