Databricks has entered the customer data platform market, announcing an agentic CDP that runs directly inside its Lakehouse architecture rather than as a separate martech stack component.

The move marks a shift for the company. Databricks built its business on data engineering and analytics infrastructure. Now it is shipping customer-facing marketing tooling of its own.

What is an "agentic CDP"?

A customer data platform (CDP) traditionally unifies customer records from multiple sources and pushes segments to marketing channels. The "agentic" label signals that Databricks wants AI agents to do part of that work — orchestrating audience building, activation and campaign logic with less manual configuration.

Placing the CDP inside the Lakehouse means customer data stays where it already lives. Marketers no longer need to copy or sync data into a standalone CDP, which has long been a cost and governance pain point in enterprise stacks.

Who does this target?

The product clearly aims at existing Databricks customers — large enterprises already running customer data on the platform. For them, the pitch is consolidation: fewer vendors, fewer data pipelines, one governance layer.

It also puts Databricks in direct competition with established CDP vendors such as Adobe, Salesforce and Tealium, which have spent years selling dedicated customer data platforms to the same enterprise buyers.

What we don't know yet

Key details remain unconfirmed at this stage:

• Pricing and packaging for the embedded CDP
• General availability timing
• Which activation channels and integrations ship at launch
• How the agentic layer handles campaign automation in practice

We will update this story as Databricks publishes specifics.

Why it matters

For marketing operations teams, the announcement signals further blurring between the data warehouse and the martech stack. If the CDP function moves into the data platform, budget and ownership questions follow — CMOs and CIOs will have to decide who buys what.

For Databricks, it is a land-and-expand play: the company can sell more software to customers already committed to its platform, while pushing AI agents as the interface layer.

Skeptics will note that "agentic" is doing heavy lifting in this pitch. The proof will be in shipped functionality — whether the agents actually reduce marketer workload, or simply add an AI label to what CDPs already did.