What is an embedded customer data platform?

Summary

• An embedded customer data platform (CDP) is built natively into the data platform where customer data already lives, rather than being a separate system that receives copies of it. Databricks describes CustomerLake as "a new Agentic CDP natively embedded in the Databricks Lakehouse."
• The CDP Institute defines a CDP as "software that creates and maintains a persistent, unified customer record that is accessible to other systems." It says CDP services can be delivered natively within the CDP or composed from external systems such as data warehouses.
• CustomerLake uses Profile Agents to build Customer 360 profiles with Agentic Identity Resolution, and Campaign Agents to build audiences, recommend next-best actions and activate across channels.
• Governance stays in Unity Catalog, and Lakehouse Federation lets teams reach customer data where it resides.
• Databricks announced CustomerLake on June 16, 2026 in Private Preview, so confirm current availability with your Databricks account team.

What is an embedded customer data platform?

An embedded customer data platform (CDP) is a CDP built natively into the data platform where customer data already lives, rather than a separate system that receives copies of it. That is how Databricks uses the term for CustomerLake, which its launch blog calls "a new Agentic CDP natively embedded in the Databricks Lakehouse." The CDP Institute defines a CDP in general as "software that creates and maintains a persistent, unified customer record that is accessible to other systems." The word "embedded" describes where that record, and the services around it, run.

What does a CDP do, and how does CustomerLake cover it?

The CDP Institute lists customer data services that include data storage, identity resolution, enrichment, governance and activation enablement. It says they may be "delivered natively within the CDP or composed from external systems such as data warehouses and other third-party components." This table maps those services to what Databricks says CustomerLake does.

CDP service	CustomerLake, as Databricks describes it
Unified profile	A governed Customer 360 that unifies behavior, transactions, engagement, product usage and operational data directly in Databricks
Identity resolution	Agentic Identity Resolution, which combines deterministic, probabilistic and agentic workflows to unify disconnected records
Enrichment	Profile Agents prepare data, identify quality issues and support third-party data enrichment
Audiences	Segments created and refined in natural language with Genie
Journeys	Campaign Agents build audiences, recommend next-best actions, activate across channels and continuously optimize experiences
Activation	Native integrations and reverse ETL to marketing tools, advertising platforms and customer engagement channels
Governance	Unity Catalog control over access, lineage and auditability

Where does the customer data live in an embedded CDP?

The CDP Institute says CDP services are enabled through "packaged, warehouse-native ('composable'), and dual-mode architectures." An embedded design comes down to data location. Databricks says teams can "build and activate customer intelligence on the governed foundation their business already uses, without copying sensitive data into a separate CDP."
Unity Catalog is "the unified governance layer for data and AI built into Databricks." Its documentation describes enforcing access control, tracking lineage and logging activity for auditing. Databricks adds that because CustomerLake is embedded in Databricks and governed by Unity Catalog, "it remains interoperable across the enterprise data estate." Customer data that sits outside Databricks stays reachable through Lakehouse Federation, which provides "governed, read-only access to external data through Unity Catalog foreign catalogs."

What should you check when evaluating an embedded CDP?

Start from the CDP Institute's definition: a persistent, unified customer record that is accessible to other systems. Then ask where each piece runs.

• Data location. Does the unified record live where your customer data is governed today, or in a separate copy?
• Identity resolution. Which matching methods does it use, and how are disconnected records unified into profiles?
• Governance. Which layer controls access, lineage and auditing for customer data?
• Activation. How do audiences and signals reach the marketing, advertising and engagement tools you already run?
• Lifecycle stage. Is the product generally available, or in preview?

What is the status of CustomerLake?

Databricks announced CustomerLake on June 16, 2026 at Data + AI Summit and said it was available in Private Preview. Treat it as a preview product and confirm current availability and scope with your Databricks account team before you plan around it.

FAQs

Does my customer data have to be in Databricks?

Not necessarily. Databricks says that through Lakehouse Federation, teams can access customer data where it resides, including in other data platforms, cloud object storage and operational databases, with Unity Catalog providing governance.

How do audiences reach the marketing tools my team already uses?

Databricks says CustomerLake uses native integrations and reverse ETL to provide bi-directional pipelines to marketing tools, advertising platforms, third-party data providers, identity graphs and customer engagement channels. Its product page describes syncing audiences and signals from Databricks to the martech, adtech and customer experience tools teams already use.

What does "agentic" mean in an agentic CDP?

In CustomerLake it refers to agents at two points. Profile Agents prepare data and unify disconnected records into profiles, and Agentic Identity Resolution adds agentic workflows to deterministic and probabilistic matching. Campaign Agents use governed customer context to build audiences, recommend next-best actions and act across channels.

What are infinity campaigns?

Databricks describes them as "continuous, agent-driven engagement loops that analyze customer signals, decide on the next-best action, and act across channels based on real-time customer context and business goals."