Knowledge Management / AI Search Question

[PRanjan]

We are evaluating ServiceNow Knowledge Management and AI Search capabilities and would appreciate guidance from the community. 

1. Is it possible to exclude Knowledge Articles from AI Search results if they have not been updated within a defined period (e.g., the last 12 months)?
   • Either through AI Search configuration, search profiles, filters, or Knowledge governance policies.
2. Can AI Search rank Knowledge Articles based on their last updated date?
   • For example, when multiple articles match a user's query, can the search experience prioritize and return the most recently updated article first?
3. Are there any out-of-the-box features, best practices, or Zurich release enhancements that support freshness-based ranking or filtering of Knowledge Articles?

came across a PPT in SNOW Community "Knowledge Management Academy.pdf" and it mentions the below points under ML Relevancy. It doesn't mention if we can add an exclusion criteria.

With the Click action, we have told the system that a user cares about:
Title match— how well the query matches the document’s title
Content match— how well the query matches the document’s body
content
Keyword match— how well the query matches the document’s tags or
meta
Freshness— how recently the document was created and/or updated
Popularity— how often the document has been viewed or used
Child Table or Template — on which child tables or templates do users
most frequently click.

Our goal is to improve search relevance by ensuring users are presented with current and actively maintained knowledge content.

Thanks in advance for your insights.

[rob_martoncik]

Yes, Thanks for your question and I think your findings allude to the answer.  There are a couple of components here.  AI search and it's subcomponents are pretty interesting and it might be worthwhile for you to check out some of that team's Academy sessions on how calculations, growth, weighting and evaluations are done.  However, over-simplifying a bit, AI search improves over time with learnings from user interactions.   So based on the findings you posted above, the answers to your three questions:  1) AI search won't surface 'expired' articles - so many customers put in simple workflows or processes to set expirations on articles X days from it's last update to do the auto-exclusion you describe.  2&3) above you can see 'freshness', so the search mechanics use that as a weight in it's learning and ranking - updated articles are higher weighted.   Hope this helps and thanks so much for your question!