How do I build a Customer 360 with warehouse or lakehouse?

Summary

• A Customer 360 is a unified profile of each customer. Databricks' CustomerLake pages describe these as "golden Customer 360 records" created "from fragmented identity data".
• Databricks' comparison says a traditional warehouse holds structured data only, while a lakehouse holds structured, semi-structured and unstructured data in an open format.
• On a lakehouse, you ingest sources, refine them in bronze, silver and gold layers (the medallion architecture), govern them with Unity Catalog, and serve them with Genie.
• CustomerLake, Databricks' agentic customer data platform, was announced on June 16, 2026 as available in Private Preview.
• Lakeflow Pipelines runs in triggered or continuous mode, so you can choose how fresh the profile needs to be.

How do I build a Customer 360 with warehouse or lakehouse?

A Customer 360 is a unified profile of each customer, built from records that sit in many systems. The CustomerLake product page describes its identity resolution as a process that "combines deterministic, probabilistic and agentic matching to create golden Customer 360 records from fragmented identity data". Below, you choose the platform first, then follow the build steps on Databricks.

Should I use a warehouse or a lakehouse?

The Databricks lakehouse docs say "A data warehouse delivers clean, structured data for BI analytics" and "The data lakehouse combines the benefits of data lakes and data warehouses". Databricks' comparison page adds these differences between a traditional warehouse and a lakehouse:

Aspect	Traditional data warehouse	Lakehouse
Types of data	Structured data only	Structured, semi-structured and unstructured (raw) data
Format	Closed, proprietary format	Open format
Reliability	High quality, reliable data	High quality, reliable data

This is Databricks' own comparison. If your Customer 360 needs only structured tables for BI, the warehouse description fits. If it also needs semi-structured or unstructured data next to those tables, the comparison points to a lakehouse.

How do I build it on the lakehouse?

1. Ingest every source. "At the ingestion layer, batch or streaming data arrives from a variety of sources." Lakeflow Connect "organizes connectors by source type, including database, SaaS, file, and streaming connectors, along with community and custom connectors for sources without a built-in option." Confirm a connector exists for each of your sources.
2. Refine in layers. "The medallion architecture describes a series of data layers that denote the quality of data stored in the lakehouse." Bronze "maintains the raw state of the data source in its original formats", silver "represents validated, cleaned, and enriched versions of the data", and gold holds "aggregated data tailored for analytics and reporting". One way to apply it: land raw CRM, web and transaction records in bronze, clean and match customer records in silver, and publish profile tables in gold.
3. Resolve identity. Matching is the step that turns fragmented records into one profile. In CustomerLake, Agentic Identity Resolution handles it with "deterministic, probabilistic and agentic matching".
4. Govern customer data. Unity Catalog is "a unified, fine-grained governance solution for data and AI". "Row filters restrict which rows a user can see in a table." "Column masks control what values a user sees for specific columns."
5. Serve and explore. Genie lets people "Ask data questions in natural language and get answers grounded in your organization's data." Genie Agents are "Domain-specific environments where data teams configure trusted data, metrics, and business rules that power Genie One answers".
6. Activate. CustomerLake can "Sync audiences and signals directly from Databricks to the martech, adtech and customer experience tools your teams already use".

What is CustomerLake?

The Databricks press release dated June 16, 2026 says "CustomerLake is now available in Private Preview." It describes an "agentic CDP built natively on Databricks". The launch blog says it works "without copying sensitive data into a separate CDP or proprietary application", and the product page says it is "governed through Databricks and Unity Catalog".
The blog describes two agent types. Profile Agents "Turn raw customer data into business-ready Customer 360 profiles directly in Databricks". Campaign Agents help marketers "build audiences, recommend next-best actions, activate across channels, and continuously optimize experiences around business goals". The blog tells interested customers to "contact your Databricks account team".

FAQs

Does a Customer 360 have to be real time?

That depends on your use case. Lakeflow Pipelines offers two modes. In triggered mode, "the system stops after refreshing all tables based on the data available when the update started." In continuous mode, "it processes new data as it arrives in data sources to keep tables throughout the pipeline fresh."

Can other engines read my Customer 360 tables?

Databricks documents Iceberg reads (UniForm): "Iceberg reads configure Delta Lake tables to automatically generate Iceberg metadata, so Iceberg clients can read Delta Lake data without rewriting files." The docs add: "A single copy of the data files supports both Delta and Iceberg clients." Iceberg client support is read-only, so writes are not supported.

Do I need a separate CDP?

Databricks positions CustomerLake as a way to build and activate customer intelligence "without copying sensitive data into a separate CDP or proprietary application". It is in Private Preview, so confirm availability with your Databricks account team before you plan around it.