Use Einstein Studio model predictions

Updated Nov 8, 2026 · Based on web research updated 2026-11-08

Einstein Studio is Data Cloud’s model-building hub: build a predictive model from scratch, or bring one in from Amazon SageMaker, Google Vertex AI, or another platform, without moving your data. Feed what it predicts — for example a churn or propensity score — into a recommender or decision so Personalization reacts to the model’s output.

Bring a prediction into Personalization

1. In Data Cloud, open Einstein Studio’s Model Builder and choose Create a model from scratch, or connect an existing model from SageMaker, Vertex AI, or another platform.
2. Train (or import) the model, then activate it against your Data Cloud data — the zero-ETL approach means no data export/import step.
3. Publish the model’s prediction output to a field on the profile or item data graph.
4. Reference that field in a recommender filter or a decision’s targeting rule.
5. Open the model’s record page and check its Prediction Jobs log to confirm predictions are still running and fresh.