Predictive Insights brings machine learning predictions directly to your semantic models. Reference an AI model, map your fields as inputs, and generate a prediction for each record in place without exporting data, switching tools, or writing code. Optionally, surface the top predictors and recommendations behind each prediction so your users can see what's driving a result and what to do next.
Where: This change applies to Data Cloud in Enterprise and Unlimited editions.
When: This update is available starting in August 2026.
Why: Predictions have typically lived in a separate data science tool, apart from the metrics and dashboards teams use daily. Tableau Semantics previously supported only built-in time-series forecasting. The PREDICT function closes that gap: it reads any predictive model registered in AI models, including bring-your-own models, so a predicted value becomes a standard field you can chart, filter, and reuse. Predictions stay consistent across every app and dashboard that reads from your semantic model.
How: Before you add predictive insights, enable this feature in any of these locations:
- Data Cloud Setup > Feature Manager
- Data Cloud Setup > Tableau Next Features
- Tableau Next workspace settings
In a semantic model, choose New > Predictive Insights. Select an AI model, map your fields to its inputs, and save. Tableau Semantics generates a prediction for each record and adds the predicted value as a field under Predictive Insights.
To return the drivers and suggested actions behind each prediction, set the number of top predictors and recommendations. When the AI model defines top predictors and recommendations, the field also returns Insight, Insight Type, and Impact Score.