CustomerLake Profile Agents: Partner integration framework
CustomerLake Profile Agents is under active development and is not yet in private preview. This page describes the direction of the Databricks partner integration model so identity, enrichment, and data hygiene providers can understand where they fit. The integration framework is still being defined, and nothing here is a finalized specification or a commitment to a particular interface, mechanism, or timeline.
What CustomerLake Profile Agents is
CustomerLake is Databricks' agentic customer data platform (CDP), embedded directly in the lakehouse. It spans two products: Profile Agents for Customer 360 and identity resolution, and Campaign Agents for segmentation, reverse ETL, and agentic personalization.
This page covers identity-focused integrations with Profile Agents. Profile Agents helps Databricks customers turn raw customer data into a well-structured Customer 360 gold layer, ready for analytics and activation. It provides an end-to-end, Genie-driven workflow across ingestion, normalization, identity resolution, clustering, survivorship, and publishing use-case-specific data products.
Where partners fit
A third-party identity provider can plug into the Profile Agents pipeline at four points. A partner may serve one or several of these.
| Stage | What the partner provides |
|---|---|
| Pre-resolution data hygiene | Standardize and cleanse PII (address standardization, contact completion, validation) before identity resolution runs, which improves downstream match quality. |
| Identity graph resolution | Match the customer's first-party records against a proprietary identity graph and assign a persistent person or household identifier. This can anchor or boost Profile Agents' own resolution. |
| Post-resolution enrichment | Once a golden record is resolved, append attributes such as demographic, psychographic, and purchase data keyed off the resolved identity. |
| Activation | Send resolved and enriched audiences to the partner's activation platform as a reverse ETL destination. The partner then distributes them across its connected channels (paid social, programmatic, CTV, retail media, publishers) using its own ID graph and channel integrations. Activation runs through CustomerLake Campaign Agents rather than Profile Agents. |
Profile Agents starts with the customer's first-party data to produce a unified profile. Partner capabilities enrich and extend that profile with third-party identity and assets.
Who this is for
This framework is aimed at partners that provide the following services:
- Data hygiene and cleansing providers that normalize, validate, and complete contact data ahead of resolution.
- Identity graph providers with a proprietary consumer, household, or business/organization graph and a persistent identifier customers can resolve against.
- Data enrichment and append providers offering demographic, psychographic, behavioral attributes, and similar data that attach to a resolved profile.
How integration works
Databricks intends to surface partner capabilities through a Databricks Marketplace Identity Hub: a curated set of partner identity, enrichment, and hygiene offerings that a customer can discover and adopt from within the CustomerLake experience.
The underlying mechanism is still under discussion. It is likely to build on Databricks' existing data collaboration primitives so that customer data stays in the lakehouse and under Unity Catalog governance. The two primitives most relevant today are:
- Clean Rooms, for privacy-preserving matching where neither side exposes its raw data.
- Delta Sharing, for sharing and querying data in place across organizations.
The exact packaging and technical interface design for Identity Marketplace is in progress. If you already integrate with Databricks through Clean Rooms, Delta Sharing, or Marketplace, you will likely be able to use those existing investments.
Get involved
Complete the CustomerLake CDP Partner Intake Form so we can understand your integration goals and partnership alignment.
What's next
- CustomerLake overview - both products and intake
- Campaign Agents - reverse ETL destinations and closed-loop measurement
- Clean Rooms - privacy-preserving matching
- Delta Sharing - share and query data in place