---
schema_version: "1"
id: 0b669758bd649d7bae7347f622eec71cef28d626350ac26224807a16ce8bb02b
title: Track Prediction Drift for Predict Jobs and Batch Transforms
collection: admin
version: current
locale: en-us
portal_url: https://docs.salesforce.com/admin/current/en-us/release-notes/264-0-0/rn_cdp_2026_winter_predictive_model_drift_integrations.md
canonical_url: https://help.salesforce.com/s/articleView?id=release-notes.rn_cdp_2026_winter_predictive_model_drift_integrations.htm&release=264.0.0&type=5
source_site: help.salesforce.com
content_type: documentation
content_hash: 469f0a901a7806952445257011f14c08a4f0360889e02f6db2ba584df8038a91
description: Monitor prediction drift and variable drift for models used in predict jobs and batch transforms. Drift charts appear in the same model monitoring interface used for live inferences, so you can spot accuracy risks without switching views. Previously, AI models monitored drift only for live inferences made through flows and APIs. Now, Data 360 compares incoming job data against reference training data to identify potential accuracy issues early, before they affect business decisions.
products:
  - Cross Cloud
release: "264"
---
# Track Prediction Drift for Predict Jobs and Batch Transforms

Monitor prediction drift and variable drift for models used in predict jobs and batch transforms. Drift charts appear in the same model monitoring interface used for live inferences, so you can spot accuracy risks without switching views. Previously, AI models monitored drift only for live inferences made through flows and APIs. Now, Data 360 compares incoming job data against reference training data to identify potential accuracy issues early, before they affect business decisions.

**Where:** This change applies to Lightning Experience in Enterprise, Performance, Unlimited, and Developer editions with Data 360.

**When:** This functionality is available starting in October 2026.

**How:** Turn on drift monitoring from **Feature Manager** in Setup. Then in Data Cloud (Data 360), on the **AI Models** tab, open a model and enable drift monitoring at the model level on the **Settings** tab. Review the model's prediction and variable drift charts to assess whether your model is performing as expected. If model drift is significant, retrain the model with new data that includes the factors contributing to the drift.