Databricks Launches CustomerLake, an Agentic CDP Embedded in the Lakehouse

Databricks has officially entered the marketing technology sector with the launch of CustomerLake, an agentic Customer Data Platform (CDP) integrated directly into its lakehouse architecture. Announced at the Data + AI Summit, the platform aims to eliminate data silos and duplication by housing customer data, AI models, and execution agents within a single governed environment. This move signals a significant shift toward composable and embedded data solutions, challenging the dominance of standalone CDP vendors in the enterprise market.

Databricks introduced CustomerLake at its Data + AI Summit in San Francisco, marking a strategic expansion into the CDP market. The platform is built natively on the Databricks lakehouse, providing core capabilities such as Customer 360 profile building, identity resolution, and real-time campaign activation. By utilizing the Unity Catalog for governance and supporting Lakehouse Federation, CustomerLake allows organizations to query data across various environments—including Snowflake and Google BigQuery—without creating new silos or duplicating data. Early adopters of the technology include major global brands such as HP, Circle K, AB InBev, and Getnet by Santander.

The product's core innovation lies in its "agentic" approach, featuring Profile Agents and Campaign Agents designed to power what the company calls "infinity campaigns." According to Databricks CEO Ali Ghodsi, these agents replace traditional manual workflows by continuously analyzing customer signals and executing next-best actions against specific business goals, such as increasing loyalty enrollment or reactivating lapsed users. This architecture is intended to solve what Databricks describes as a fundamental flaw in legacy CDPs, which typically require data to be moved into external systems, thereby slowing down marketing teams and creating security risks.

Industry analysts view this launch as a disruptive force that could redefine the enterprise platform landscape. David Raab, founder of the CDP Institute, noted that the move simplifies operations for IT and marketing teams while posing a direct challenge to conventional CDP vendors. Gartner further supports this trend, predicting that 80% of new enterprise CDP deployments will be embedded or composable by 2030. To support this ecosystem, Databricks has partnered with 21 launch partners, including Bloomreach, which is integrating its Loomi marketing agent with CustomerLake to enable personalized cross-channel execution without manual data handoffs.