Knowledge Article Keywords - Best Practices

[mwsteubinge]

Hello,

I recently started a new role and one of my main responsibilities is managing and enhancing my organization's Knowledge Library. I want to enhance our users' experience and ability to find what they need in the service portal by making sure all of our Knowledge Articles include keywords that will make them populate in corresponding searches. I've struggled to find any recent guidance on best practices for keyword formatting. The most recent thing I can find is from this post that is over five years old: https://www.servicenow.com/community/servicenow-ai-platform-forum/keyword-best-practices-for-knowled...

Adding keywords

An article's keywords field should include all the relevant terms your user might try to find your article. 

Put yourself in the user's shoes. How would they describe the task they're trying to complete? What words do they use when talking about this topic? The language you use to describe your service could be very different than the language your customer uses, so remember to include both terms in the keywords.

• Separate terms using commas (e.g. search, SEO, keywords)
• Separate words in the same phrase using underscores (e.g. how_do_I_edit_knowledge_keywords, improve_knowledge_search)
• Avoid generic terms that will pull up too many results. If all authors add the keyword "article" to every KB, it becomes meaningless.
• List questions users might have that could be answered by this article. This greatly reduces the number of FAQ articles you need, while providing a better user experience. 
  • Remember to include different variations of the same questions and phrases (e.g. "How do I edit a knowledge article" and "How to edit a knowledge article" and "How do you edit a knowledge article" and "how to edit a KB") Each question should be included.
• Repeat keywords up to three times to boost relevancy.

Can anyone confirm that the above guidance is still the best way to approach Knowledge Article keywords?

Thank you for your help!

[Leri Andrews]

I think the guidance you have posted worked well for Zing search where one could apply a weighting to the meta field.  With the recent advances firstly with AI search and then the generative Now Assist, I believe the meta field is less important.  I would still separate with commas and avoid superfluous words like 'policy' but not spend so much time agonising over question phrases and repeating words.  There's also a lack of clarity over how the synonym dictionary works with the meta so, for example, in an article about 'What sort of jumper suits your body type' is it more important to have 'sweater' and 'pullover' in the meta or the synonym dictionary? 🙂  I'd suggest making sure the article short-description and body refer to these words to be on the safe side!!