Discovering knowledge articles and catalog items for employees

  • Release version: Washingtondc
  • Updated February 1, 2024
  • 1 minute to read

    • Discover all the users who have a similar profile as the employee who has logged in to Employee Service Center. Identify the top viewed knowledge articles and top submitted catalog items by those users. Display the top three knowledge articles and top three catalog items to the employee.

    Top three knowledge articles and catalog items

    When the User Profile Based Recommendation (ml_sn_sn_hr_core_global_user_profile_based_recommendation) solution definition is configured and the predictive model is trained, the top three knowledge articles and catalog items are displayed to an employee in Employee Service Center and HR Ticket page.
    The top viewed knowledge articles and top submitted catalog items as displayed to an employee.
    Note:
    The number of catalog items and knowledge articles displayed in the Relevant for you widget depends on the number in the Limit field in Employee Center. By default, the Limit field is set to 3. For more information, see Modify the Recommended for you widget display.

    Auto training the predictive model

    By default, the User Profile Based Recommendation solution definition is configured and the predictive model is auto trained when all the following conditions are met:
    • The Human Resources Scoped App: Core (com.sn_hr_core) plugin is installed.
    • The Predictive Intelligence (com.glide.platform_ml) plugin is installed.
    • The glide.platform_ml.auto_training.enabled system property is set to true.
    You can manually train the predictive model if it is not auto trained by default. After you manually train the predictive model, navigate to the HR AI configuration module, open the Similar users group record, and map the configured solution definition.

    Configuring widget options

    You can also configure the following widget options:
    • User prediction count: Number of similar users discovered by the predictive model. By default, the top 25 similar users are discovered for the user.
    • Recent activity cut off days: Number of days considered by the predictive model for querying the articles and catalog items. By default, the last 30 days activity is checked from the Knowledge Use table, Requested Items table, and Items Produced Records table.
    To configure the widget options, navigate to the Service Portal Configuration module, select the widget editor, search for Relevant for you, and click Edit Option Schema.