Loading
Prepare for Email to Become the Default Login ExperienceRead More
Tableau Semantics
Add a Predictive Insight to a Semantic Model (Beta)

You are here:

  1. Salesforce Help
  2. Docs
  3. Tableau Semantics

Add a Predictive Insight to a Semantic Model (Beta)

Add a Predictive Insight to surface an AI model's prediction as a field in your semantic model. For each record your model returns, Tableau Semantics passes the mapped field values to the AI model and returns the predicted value in real time.

Required Editions

Available in: All editions supported by Data 360. See Data 360 edition availability.
User Permissions Needed
To add or edit a Predictive Insight:

Permission set:

  • Data Cloud Architect

AND

Semantic Models object
Note
Note Predictive Insights in Tableau Semantics is a pilot or beta service subject to the Beta Services Terms at Agreements - Salesforce.com or a written Unified Pilot Agreement if executed by Customer, and applicable terms in the Product Terms Directory. Use of this pilot or beta service is at the Customer's sole discretion.

Before you begin, confirm that:

  • You have turned on the Predictive Insights in Tableau Semantics feature in any of these locations in your org:

    • Data Cloud Setup > Feature Manager
    • Data Cloud Setup > Tableau Next Features
    • Tableau Next workspace settings
  • You have edit access to the semantic model.
  • At least one AI model is available in Data 360. Any Data Cloud license includes read access to AI Models. If no models are registered, the New > Predictive Insights option appears but you can't select a model. See Using AI Models in Data 360.

For background on how Predictive Insights work and what they return, see Predictive Insights in Tableau Semantics (Beta).

  1. In Tableau Semantics, open the semantic model you want to edit.
  2. On the model canvas, select New, and then select Predictive Insights.
    Select Predictive Insights for Model
  3. In the Predictive Insights dialog, select a model from the Predictive Model picker.
    Only predictive models hosted in Data 360 appear in the picker.
  4. Under Map Fields, map each of the AI model's required input fields to a corresponding field from your semantic model.
    • Inputs can be dimensions, measures, or calculated fields. Pass a calculation—for example, SUM([Sales])—instead of a raw field.
    • The matching field name from the model is preselected by default. Select a different field to change it.
    • Map all fields for the most accurate predictions.
    • Don't use a Predictive Insight field as an input to another Predictive Insight.
      View Predictive Insights
  5. (Optional) To surface the drivers and suggested actions behind each prediction, under Options, set Number of Top Predictors and Number of Recommendations to a value greater than 0.
    Higher values increase processing time. Keep these values as low as your use case requires. Leave both at 0 if you don't need explainability.
  6. Enter a Predictive Insight Name and confirm the auto-generated Predictive Insight API Name. Optionally, add a description.
    Predictive Insights Name and API Name
  7. Save your changes.
    View Predictive Insights
    The Predictive Insight appears under Predictive Insights in the left panel. When top predictors or recommendations are turned on, the field also produces Insight, Insight Type, and Impact Score columns.

To edit or remove a predictive insight, expand Predictive Insights in the left panel, select the options menu next to the insight, and then select Edit or Remove. Editing an AI model doesn't break the Predictive Insights that use it. If an edit removes an input field from the AI model, the remaining fields stay mapped and the Predictive Insight continues to work.

 
Loading
Salesforce Help | Article