How to Make Service Portal Search Return Results Regardless of User Language?

[Geraldina Valer]

Hi everyone,
I need some help understanding how to configure the Service Portal search behavior.

Right now, when users type a search term (for example "base de conocimiento"), the portal only returns results that match content in the same language the user is currently logged in with.

What I need is different:
I want the search to return all matching records, such as Knowledge Articles and Catalog Items, regardless of the language of the logged-in user.

For example, if a user searches for "base de conocimiento", I would like the results to also include items or articles that contain English terms like "knowledge base" or "knowledge article", even if the user is logged in in Spanish.

Is there a way to configure the Service Portal search so that it ignores the user’s language and retrieves all relevant results across all languages (even if that means having duplicates)?

Thanks in advance!

[gennaropaga]

Out of the box, Service Portal search does not work the way you’re describing.

Search in Service Portal is language-aware by design. It uses the user’s session language and the indexed translations, so if a user is logged in in Spanish, the search will only return content indexed for Spanish. It will not automatically match Spanish terms like “base de conocimiento” with English content such as “knowledge base” or “knowledge article”.

There is no standard setting or property to “ignore the user language” or to force the search to return results across all languages. Dynamic translation and Knowledge language configuration do not affect search behavior, and Zing synonyms are language-specific, so they don’t solve cross-language matching either.

If you need this behavior, the only options are custom or licensed features:

– You can build a custom search source (or override the existing one) and query the underlying tables without filtering by language. This works, but it’s a full customization and you need to be careful with performance, especially with large knowledge bases.

– A more common workaround is to expand the search terms yourself. For example, when the user searches for “base de conocimiento”, you programmatically expand the query to also include “knowledge base” and similar English terms. This requires maintaining a multilingual synonym map, but it’s predictable and relatively simple to control.

– If you have AI Search / Now Assist, semantic and cross-language matching may be possible, but this depends on licensing and needs to be validated on your instance.

Thanks.