---
schema_version: "1"
id: 4ebf95b36bfa13c80f0b226544968d38d0e81316745b5efd54289c126b236ced
title: Discover Patterns in Your Data Without Training a Clustering Model
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_clustering_reusable.md
canonical_url: https://help.salesforce.com/s/articleView?id=release-notes.rn_cdp_2026_winter_predictive_model_clustering_reusable.htm&release=264.0.0&type=5
source_site: help.salesforce.com
content_type: documentation
content_hash: 166c65a47a98a84f39c3777b753d0513963a93ca94b6855269a2aedabcf729ef
description: Dynamic clustering in AI Models uncovers patterns and creates targeted customer segments by grouping similar records, such as customer profiles. A built-in algorithm assigns records to clusters each time you run a predict job, requiring no training data or reusable model.
products:
  - Cross Cloud
release: "264"
---
# Discover Patterns in Your Data Without Training a Clustering Model

Dynamic clustering in AI Models uncovers patterns and creates targeted customer segments by grouping similar records, such as customer profiles. A built-in algorithm assigns records to clusters each time you run a predict job, requiring no training data or reusable model.

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

**When:** This feature is available starting in September 2026.

**Why:** Analyze harmonized data to uncover hidden patterns and relationships. Use dynamic cluster assignments to create targeted customer segments. Enrich records with variables for downstream AI models. You can also organize large datasets into manageable clusters for forecasting or multiclass models.

**How:** In Data Cloud, go to the AI Models tab and select **Cluster**. Select either K-Means Clustering or Density-Based Spatial Clustering (HDBSCAN), and then run the predict job. The output includes a cluster ID for each record. Because assignments are generated dynamically, cluster IDs can change when your source data or model configuration changes. This release doesn't support human-readable cluster labels.