Clarification on AI Search matching behavior in Knowledge Portal

Geraldina Valer
Tera Contributor

4 hours ago

Hi,

I need support to understand how the search engine is matching keywords in the Service Portal (or Employee Center).

I created a Knowledge article with the following fields:

  • Short description: “fantasia”

  • Body: “fantasma”

  • Metadata: “magia”

When I search for “fantasia”, the result list includes not only my article, but also unrelated articles.
For example, in the screenshot attached, the search shows: “Showing 4 results for camtasia”.

My question is:

How is “fantasia” being matched with “camtasia”?
I would like to understand which part of the AI Search / indexing process is generating this match

Could you please provide guidance or documentation on how the search engine maps keywords and why it returns these kinds of approximate matches?

Thanks and regards,
Geraldina

Preview file
61 KB
Preview file
102 KB
0 Helpfuls
  • 120 Views
1 REPLY 1

k_lutz
Giga Sage

an hour ago

Hi @Geraldina Valer 

 

The first suggestion I would have...is when you are testing like that I always run the knowledge indexing after adding new knowledge articles I am going to test. This way I know it has run through that process and will be properly picked up by the search since the whole Search Application > Search Profile > Search Source > Indexed source can get to be sooo much fun to make sure all things are in place.

 

The second part I would have to say sounds like what Kristen Dettman posted about "fuzzy" searches: 

https://www.linkedin.com/posts/kristendettman_fuzzy-search-in-servicenow-helpful-or-chaotic-activity...

0 Helpfuls