How does ServiceNow Knowledge search rank articles, and does the Meta field improve searchability?

KrystynaG
Giga Contributor

3 weeks ago

Hi everyone,

I’m looking to understand how the ServiceNow Knowledge search engine works, particularly how it matches a user’s query to knowledge articles and ranks the results.
Which article fields influence search relevance- is it the title, keywords in the body or metadata?

Does adding relevant terms to the Meta field improve an article’s searchability or ranking?
If so what are the most recent recommendations/ best practises for using it?
Also will answer differs between standard Knowledge search and AI Search?

Thanks a lot 

0 Helpfuls
  • 687 Views
4 REPLIES 4

Aditya_hublikar
Giga Sage

3 weeks ago - last edited 3 weeks ago

Hello @KrystynaG ,

In simple words

  • Zing: Higher ts_weight can make matching fields more influential in search ranking.

  • AI Search: Field configuration and relevance tuning can affect which articles rank higher.

  • Search location/context: Determines which articles are eligible to appear, based on factors such as knowledge base, access permissions and search configuration.

You can refer this for more details:

Weight of articles search  

 

If this helps you then mark it as helpful and accept as solution.

regards,

aditya

 

0 Helpfuls

MaryG
ServiceNow Employee

3 weeks ago

Great question, and an important distinction between Zing and AI Search.

For legacy Zing search, ServiceNow documents that both the Short description and Meta fields on kb_knowledge have a default ts_weight of 10. This means terms in those fields can influence relevance scoring, along with other factors such as term frequency and proximity.

AI Search operates differently. Searchability depends on the configured indexed source and field settings, while ranking is influenced by the AI Search relevance configuration. We should not assume that Zing field weights carry over to AI Search.

From a Knowledge Management perspective, my recommendation is to:

  • Write clear, descriptive article titles using terminology employees actually search for.

  • Include relevant terminology and alternative expressions naturally in the article content.

  • Use the Meta field for meaningful supplemental keywords, rather than repeating terms or attempting to manipulate ranking.

  • Validate search results in the actual search experience being used, whether Knowledge search, Employee Center, or another configured experience.

The broader principle is that metadata can support discoverability, but it should never compensate for poorly structured or incomplete article content.

References:

Set the relative weight of a field

AI Search indexed sources

drosenst2
Tera Contributor

2 weeks ago

I strongly suggest using meta. It's great for public articles that are open to different countries. For instance, I just completed a whole meta project for Canada. They spell the same word slightly different i.e. centre, behaviour, etc. Once we did that, it made it much easier for them to find their material.