HR PIWB template: Auto-categorize email cases

  • Freigeben Version: Australia
  • Aktualisiert 12. März 2026
  • 4 Minuten Lesedauer

    • Use a guided template that walks you through setting up a machine learning model to categorize the email cases automatically for improved productivity and cost savings.

    Vorbereitungen

    Role required: sn_piwb_hr_content.admin

    Warum und wann dieser Vorgang ausgeführt wird

    This template walks you through customizing a use case model to categorize the email cases. When the use case template shows the label Guided, you can use several implementation steps and are automatically taken to evaluate and tune your models when you click Start. Otherwise, you begin by creating a machine learning model.

    Prozedur

    1. Navigate to All > Predictive Intelligence Workbench > Use Cases > Create New from Templates.
    2. Select the Email case categorization for HR guided template.
      Abbildung : 1. Email case categorization use case
      PIWB Email case categorization use case model
    3. Provide a unique name for your use case in the Use case name field.
      Hinweis:
      The use case contains multiple models that you create. You can use the same name as the template, if required.
    4. Provide a name for the model in the Model name field.
      Hinweis:
      Models are trained based on default parameters. The name of the model should reflect its purpose, for example, Email case categorization for HR.
    5. Click Start.
      Hinweis:
      To start implementation, you must provide a use case name and short description.
      The use case setup page opens displaying the name and description of the use case you created. On this page, you can see all the implementation phases to create and implement your use case model.
    6. In the Create and train models section of the setup, click Start to create a model associated with your use case.
    7. Fill in the required fields with a Model name and Short description.
      Hinweis:
      Other data related to the model is pre-filled by default, such as the Data table, Predicted field, and Processing Language.
    8. Expand the Review the filters used to train this model section to view the default filters.

      Review the base system parameters set to help train the model in the Input fields and Number of records fields. To modify this data, click Advanced Setup and then Save any changes you make. You can customize the filters to best represent your business data or add new criteria by clicking New Criteria.

      Hinweis:
      By clicking Advanced Setup you can change the processing language, review, and modify the filters used to train the model. You can customize the filters to best represent your business data or add new criteria by clicking New Criteria. Review and modify the input fields used to generate predictions. Customize the fields to best represent your business data by moving fields between the Available and Selected lists. Ensure that you have sufficient number of records for the model. Save any changes you make.
    9. Click Save.
      The use case ML Model setup page opens.
    10. Click Continue in the Create and train models section.
      The Create a model page opens.
    11. Click Train this model.
      The process may take a while. The Train this model window appears letting you know that the process takes a while.
    12. Click Start to initiate training.
      The ML Model setup page opens.
    13. Click View Progress below the header on the ML Model setup page to monitor the training process and click close.
    14. Wahlweise: Click View Progress below the header on the use case setup page to monitor the training process.
    15. Click Start to initiate training.
      The use case setup page opens with the trained model.
    16. On the available trained model, click the Tune values.
      The use case model record opens. You can view the Precision(%), Coverage(%), and Net Automation(%) scores.
      Hinweis:
      Refine your use case model by defining the right combination of coverage and precision values.
    17. When you are ready to test your use case model, click Test your models .
      Abbildung : 2. Guided Steps
      PIWB model guided steps for training, testing, and integration
      The use case Testing your models page opens.
    18. In the Select models to test section, refine the models for testing and comparison.
      Move the options from Available to the Selected. Remember, selecting more models takes more processing time for a batch test.
    19. In the Define testing parameters section, decide if you want to test one use case model.

      Select the Single test test type. Single test is the default.

      Hinweis:
      Select Batch test when you want to test more than one use case model.

      Determine the number of top results you want to display.

    20. In the Input fields section, provide a short description of your use case model test.
    21. Click Run Test.
      View the test results data in the View test results section with details such as predicted values and confidence.
    22. Click Mark as complete to complete testing.
    23. When you are ready to integrate your use case model into your business processes, click Start in the Integrate a model section.
    24. In the Select a model to integrate section, select the model from the available list.
    25. In the Retraining Schedule field you can change the definition, if desired.
      The default value is Run Once, but you can retrain as often as every 30 days to every 180 days.
    26. Click Integrate.
      On the confirmation pop-up, click Integrate again to complete this action. You have integrated a use case model into your business process.

    Ergebnisse

    When the configuration is complete, the solution auto-categorizes the HR service for email cases. For more information, see Auto-case creation from an email

    Nächste Maßnahme

    You can verify the integration status from HR Administration > HR AI Configurations > Solution definition. The use case is now mapped to the selected solution definition.