NLU bringing very wrong results with 100% reliability.

Albergaria
Tera Contributor

We have identified strange results using NLU. Some expressions that are not registered in the intent are achieved with 100% reliability. In addition, by using random letters we also hit some intentions the wrong way. Even giving negative feedback on these predictions, they keep showing up. We would like to know how to avoid these wrong results appearing in user searches.

 

I looked up "llllllllllll" and it found an intent with 100% reliability and no utterance of that intent has anything remotely like what I searched for:

VictorAlbergar_0-1680012828629.png

 

 

20 REPLIES 20

Brian Bakker
ServiceNow Employee
ServiceNow Employee

@k_lutz 

Please raise a case with ServiceNow support and ask for Brian Bakker to review your NLU Model, and I will review and fix any issues with it.

Brian Bakker
ServiceNow Employee
ServiceNow Employee

@k_lutz 

I am sorry that the support you received on this issue has not been satisfactory. Please can you message me the case ID, and I will pick it up from the assigned TSE and I will assist you further to understand the root cause to this issue and how to resolve it.

Thanks Brian. I will let you know next week. We are cloning production down to several of our non-prods this weekend and support is going to meet with me next week so I can show them how I go through the steps you mention as well as the ones I have seen in other postings and the system continues to ignore when the "No intents in this model match this utterance" is used. I also added the synonym like you suggested.

 

 

Hi Brian,

Just wanted to let you know that I was able to get this working in one environment by basically deleting everything and putting it back together.

 

Although, now that I moved it to our "test" environment to see if it works...its not and getting different results. A lot if it is the OOTB models rating things even higher than the prior dev environment for things that do not even make any sense. Raising more tickets with support but really would advise against using the product(s).

Brian Bakker
ServiceNow Employee
ServiceNow Employee

@k_lutz 

Please let me know you case numbers, so that I can assist you.