Databricks has officially entered the marketing technology market with CustomerLake, an agentic Customer Data Platform (CDP) announced by the company. The move marks the lakehouse vendor's first dedicated product aimed squarely at marketers and marks a shift in how enterprise CDPs get built: on top of an open data lakehouse rather than as a packaged marketing cloud.

The announcement is short on detail but long on signal. Databricks — best known for its data engineering and AI platform used by thousands of enterprises — is now competing directly with established CDP vendors such as Adobe Real-Time CDP, Salesforce Data Cloud and Tealium, all of which have been racing to add generative and agentic AI capabilities to their own offerings.

What is an "agentic" CDP?

The word "agentic" is the operative part of the product name. Rather than serving purely as a rules-based pipe that unifies customer records and syncs them to activation channels, an agentic CDP puts AI agents to work on the data layer itself.

For practitioners, that framing suggests a platform where autonomous or semi-autonomous agents can:

• assemble and maintain unified customer profiles;
• build audiences and segments on request;
• orchestrate campaigns or journeys across channels;
• answer natural-language questions about customer data.

Databricks has not yet published full capability, pricing or integration specifics for CustomerLake in the initial announcement, so marketing teams evaluating the platform should watch for those details in follow-up documentation.

Why does Databricks want the marketing stack?

The strategic logic is straightforward. Customer data has become the most valuable — and most fragmented — dataset in the enterprise. Marketing teams have spent years fighting identity resolution problems, siloed session data and restrictive walled gardens.

Databricks already holds much of that raw customer data. Many of its enterprise customers run their transactional, behavioral and clickstream data on the Databricks Lakehouse. Launching a CDP lets the company move up the stack from infrastructure to marketing workflows — capturing budget currently spent with dedicated CDP and marketing cloud vendors.

The approach also plays to Databricks' positioning on open data formats and its years of investment in unified analytics and AI, including generative AI tooling for data workloads.

What changes for martech practitioners?

For CDP buyers and marketing operations teams, the entrance of a major data-platform vendor changes the evaluation calculus in three ways.

First, architecture. A lakehouse-native CDP means customer data does not need to be copied into a separate CDP store — long a source of cost, latency and governance headaches. Profiles and audiences can, in principle, be built directly where the data already lives.

Second, AI leverage. "Agentic" capabilities promise to reduce the dependency on technical teams for segmentation and data prep, putting more self-service power in marketers' hands.

Third, competitive pressure. A new entrant with enterprise data credibility will force incumbent CDP vendors to sharpen their own AI and openness roadmaps — which historically benefits buyers through faster feature velocity.

What to watch next

Key open questions for anyone tracking this launch:

• Pricing and packaging: Databricks has not disclosed how CustomerLake will be priced relative to its existing platform consumption model.
• Activation integrations: A CDP lives or dies on its connectors to ad platforms, email and messaging tools; the integration list will determine real-world usefulness.
• GA timing: The announcement does not specify general availability, so teams should monitor Databricks' release calendar.

What is clear today is the direction: Databricks intends to be a serious player in customer data for marketing, and it is bringing an agent-first architecture to a category still dominated by packaged marketing clouds.