CDP on Databricks

Summary

• A CDP on Databricks runs customer data platform work on the Databricks lakehouse. Databricks offers CustomerLake, its agentic CDP, and you can also build the data layer yourself with Lakeflow Connect, Lakeflow Pipelines and Unity Catalog.
• CustomerLake is in Private Preview. Databricks announced it on June 16, 2026 with the sentence "CustomerLake is now available in Private Preview" and points interested teams to their Databricks account team.
• Databricks says CustomerLake is built natively in Databricks, so teams can build and activate customer intelligence without copying sensitive data into a separate CDP or proprietary application.
• Databricks says CustomerLake is governed through Databricks and Unity Catalog, giving data and IT teams centralized control over access, lineage and auditability across customer data, audiences and AI assets.
• Databricks says CustomerLake syncs audiences and signals directly from Databricks to the martech, adtech and customer experience tools your teams already use.

CDP on Databricks

A CDP on Databricks runs the customer data platform jobs (collecting customer data, building a unified view of each customer, and activating audiences) on the Databricks lakehouse instead of in a separate system. Databricks' agentic CDP is CustomerLake, which Databricks announced in Private Preview on June 16, 2026. You can also build the data layer yourself with platform components such as Lakeflow Connect, Lakeflow Pipelines and Unity Catalog.

What is a CDP, and which Databricks capability covers each job?

Databricks' explainer "What Is a Customer Data Platform (CDP)?" says a CDP "allows marketers to create a unified view of their customers by consolidating and integrating data from multiple sources." It groups the work into data collection, data management and modeling, and data activation. The table maps those jobs to capabilities described in the Databricks documentation.

CDP job	Databricks capability	What the documentation says
Collect data	Lakeflow Connect	Managed connectors produce ingestion pipelines governed by Unity Catalog and powered by serverless compute. Connectors are in various release states.
Clean and model data	Lakeflow Pipelines	Extends Spark Declarative Pipelines, a declarative framework for batch and streaming data pipelines in SQL and Python.
Read data in place	Lakehouse Federation	Governed, read-only access to external data through Unity Catalog foreign catalogs, without migrating your data.
Control access	Unity Catalog	Privileges, attribute-based policies, and row and column filters, plus lineage and an audit log system table.
Ask questions	Genie	Users ask data questions in natural language, and answers are governed through Unity Catalog.
Resolve identities and activate audiences	CustomerLake (Private Preview)	Profile Agents, Campaign Agents, and bi-directional syncing with marketing tools.

How do the medallion layers apply to customer data?

Databricks defines the medallion architecture as "a series of data layers that denote the quality of data stored in the lakehouse." Bronze keeps the raw state of each source, silver is where you perform data cleansing, deduplication, and normalization, and gold holds highly refined views of the data that drive downstream analytics. For example, you could keep raw web, CRM and transaction extracts in bronze, cleaned and deduplicated customer records in silver, and the profile and audience tables that marketing teams read in gold.

What does CustomerLake add, and what is its status?

Databricks announced CustomerLake at Data + AI Summit on June 16, 2026. The launch post states: "CustomerLake is now available in Private Preview." It points interested customers to their Databricks account team.

• Profile Agents help turn raw customer data into business-ready Customer 360 profiles directly in Databricks. Agentic Identity Resolution combines deterministic, probabilistic, and agentic workflows to unify disconnected records.
• Campaign Agents use governed context in Databricks to help identify the right audience, recommend the next-best offer or channel, activate the campaign across destinations, and optimize or suppress based on performance signals.
• The marketer workspace is powered by Genie, so marketers can ask questions and build audiences in natural language.
• Serving and models. Databricks lists a real-time serving layer for customer attributes, and says you can deploy custom models through Model Serving.

How does governance work for customer data on Databricks?

Unity Catalog is the unified governance layer for data and AI built into Databricks. You manage who can access what with privileges, attribute-based policies, and row and column filters. A row filter restricts which rows a user can see, and its filter function evaluates each row at query time. Unity Catalog also tracks lineage and keeps a record of data access in the audit log system table. Databricks says CustomerLake is governed through Databricks and Unity Catalog, with centralized control over access, lineage and auditability across customer data, audiences and AI assets.

What should you confirm before you build a CDP on Databricks?

• Which lifecycle stage you can accept. CustomerLake is in Private Preview, and Lakeflow Connect connectors are in various release states, so check each connector you need.
• Which tools receive audiences. Ask your account team whether your martech, adtech and customer experience destinations are supported.

FAQs

Which lifecycle stage is CustomerLake in?

Databricks states CustomerLake's stage as Private Preview, announced on June 16, 2026, and directs interested customers to their Databricks account team. Confirm the current stage with them before you plan a rollout.

Does CustomerLake replace my marketing tools?

Databricks describes CustomerLake as open and composable, and designed to work with the marketing, advertising, analytics, and customer experience tools teams already use. It syncs audiences and signals from Databricks to those tools.

Can I build a CDP on Databricks without CustomerLake?

Yes, for the data layer. Ingest with Lakeflow Connect, model with Lakeflow Pipelines, and govern with Unity Catalog, as the table shows. Agentic identity resolution and audience activation are documented as CustomerLake capabilities, so plan how you'll cover those two jobs if you don't use it.