Exclude topic discovery via NLU utterances?

[Keszia]

I was wanting to know if we can identify phrases that doesn't trigger a topic that may have a shared keyword or utterance, i.e., if the user says "teams meeting guide" I don't want it to trigger the topic about "meeting room issues". I know we can toggle topic discovery, but that hides it forever, unless you select it from the list.

[Abbas_5]

Hello @Keszia ,
Please refer to the below link:
https://www.servicenow.com/community/virtual-agent-nlu-articles/hide-topics-to-be-discoverable-via-k...

If it is helpful, mark it as a thumps-up and accept the correct solution.
Thanks & Regards,
Abbas

[Keszia]

Hi @Abbas_5, 

That solution only partly answers my question - as it hides the topic forever. I was wanting to know if there is a way to exclude utterances - so if I were to say "teams meeting guide" it wouldn't trigger the topic about "meeting room issues". Like how we have intents to trigger it, I was wanting the opposite - what does trigger it, vs what doesn't.

[Victor Guzman]

Hi @Keszia , 

Did you find a solution? I'm facing the same situation 😞

btw, I found a ServiceNow article explaining that topic discoverability is evaluated at runtime
https://docs.servicenow.com/bundle/vancouver-servicenow-platform/page/administer/virtual-agent/conce...

[Keszia]

Hi @Victor Guzman not yet - our temporary fix was to increase the confidence level of the Virtual Agent while we work on something more permanent. That article you linked might help - I'll have to share it with our devs and see what's possible. Thanks!