How does "Recommended for you" and "Popular topics" widgets work on Employee Center Pro?

[Aditya32]

There's "Recommended for you" and "Popular topics" widgets available on Employee Center Pro homepage.

ServiceNow product documentation says "The Recommended for you widget displays catalog items and knowledge articles, enabling you to view information that is most relevant to you.". How is ServiceNow determining what's relevant to the logged in user?

ServiceNow product documentation says "The Popular topics widget lets you quickly view and access the most viewed topics. Employee Center uses an algorithm to determine which topics must appear on the widget.". Does it show the most viewed topics by the logged in user or does it show the most viewed topics by everyone in the company?

[Øyvind Fosse · Accepted solution]

Hi Aditya,

I recommend you read this FAQ for a full understanding of the behaviour of these two widgets:

FAQs on Curated Experiences, Search Configurations, and Content Recommendation

the essential parts for your questions is the following:

3. What is the recommendation logic used with ‘Recommended for you’ widget?

In general, this widget highlights the most viewed knowledge article and most requested catalog items across all employees (for a time period defined in the instance options by the portal administrator)

For customers who have Predictive Intelligence license and have it activated in their instance, this widget will use the similarity framework applied on the Employee’s profile to make the content recommendation.

See the flowchart below:

4. What is the recommendation logic used with the ‘Popular topics’ widget?

We define "popular" topics as topics that contain content (catalog or knowledge) that have higher popularity. For a catalog item, popularity is the number of times it was requested and for a knowledge article, it is the number of times it was viewed. The popularity of a topic is the aggregation of the popularity of all the content under a topic.

Popularity is calculated based on a daily scheduled job that looks at the entire history, since the day Employee Center was turned on the instance and across all users in the instance.

[JacobReidu]

In my opinion, most recommendation systems use content-based and collaborative filtering to show user recommendations and provide a better user experience. I came across this when researching a startup called Vody, one of the authors of which is Jeremy Houghton. Well, for a recommendation engine to succeed, it has to have the ability to find the best recommendations for people, so it makes sense to focus on finding what people would like at work or home. To do this, recommendation systems should only display or provide meaningful items to the user.

[Ja1]

Hi,

If the user has a different language set to the base language, will recommendated articles and items be displayed in their selected language (if translated articles and items exist)?

I've been testing in my PDI instance and have found that this doesn't work but not sure if that's just my PDI instance or if I've missconfigured something.

[Community Alums]

'Getting similar users using the HR profile'. 
How exactly are 'similar' users defined for this widget? What HR/profile or usability data is used?

I cannot find documentation of this.