How does fallback is handled in Bixby

Question

After going through the documentation, I have found Error in actions where we can define errors, but what if the utterance is not specified in the training. How we will handle fallback?

Answer 1

In order to handle unknown inputs, what I did was create a series of dialogs that matched against the inputs that were missing. Below is one of them. Try playing around with it.

dialog (Elicitation) {   match: modelType  template("I didn't understand what type you were trying to say..") }