Is there any way that i can control the ambiguity responses from the bot with respect to any Intent or FAQ resolution against the user utterance?


(Baba Farooq Basha Renumanu) #1

Based on the score that has been returned by the NL Engine, you can always control the Bot responses with respect to Intents or FAQ’s resolution and avoid the ambiguity responses from the Bot. In order to achieve this, you can customize the NLP thresholds. You can find required settings on the below-mentioned path.

Bot -> Natural Language -> Advanced Settings -> THRESHOLDS -> Fundamental meaning Thresholds or Knowledge Collection Thresholds.


(Sunitha S) #2

Few more tips:

  1. Ensure to use unique names across dialogs to avoid ambiguity
  2. Mark relevant words as mandatory incase of ontology if there are multiple similar utterances . This will avoid ambiguity as well ex. Reset password for sap , reset password for TL etc. Here, SAP and TL can be marked as mandatory.
  3. Recheck on ML utterances as well . Perhaps , you have redundant utterances in multiple tasks?

Hope that helps