virtual agent- NLU model predicting random and inconsistent topics.

Pallavi Shetty
Tera Contributor


The virtual agent NLU model is giving conflicts with other topics which are changing every time. Resolving has become problematic. The conflict with other topics is random for different users for the same utterance.

 

Note : We have more than 1 model.

The feedback loop is also not working.

 

Problem 1: When 2 different users at the same time type the utterance “xyz”, both the users are getting different results.

Problem 2: The conflicts are different every time and also the model is predicting different in Teams and Portal.

Problem 3: Even though the intent does not contain utterances the model is predicting those topics.

 

We would like to know how to avoid these wrong results appearing in user searches.

0 REPLIES 0