Databricks ended weeks of rumors today as it announced CustomerLake, an agent CDP.
The announcement came at the company’s Data + AI Conference in San Francisco and marks Databricks’ move into the martech market. In March, Databricks entered the security market with the announcement of its Lakewatch product.
CustomerLake provides marketers and data teams with agents that analyze behavior, make decisions, and act. The company says these employees can deliver personalized customer experiences 1 billion times a day.
Built on Databricks lakehouse technology and powered by Unity Catalog, CustomerLake integrates customer data, identity optimization, audience building, campaign automation, and activation.
With the announcement of CustomerLake, Databricks is attracting marketing organizations facing a future where agents are used internally by salespeople, while at the same time marketers will have to market to agents sent by customers to research and test products.
As Databricks sees it, most traditional martech applications are not built with any scenario in mind.
Does the agent era require a new type of CDP?
Value CDPs follow the waterfall model: campaigns are planned and executed across disconnected systems, can take weeks to deploy, and leave customer data locked out of the company’s core AI platform. The result is a broken identity that makes true personalization impossible at scale.
The agent era, on the other hand, requires real-time access to context, data, and execution. With CustomerLake, Databricks brings CDP natively to the Databricks environment, so that the same models that generate insights can directly drive performance.
“Marketers need to rethink their entire foundation — not just the campaigns they run, but the customers they run, now including agents,” said Ali Ghodsi, founder and CEO of Databricks, in a statement. “With CustomerLake, customer data, AI models, and agents live in one central control. Marketing ceases to be a series of campaigns and becomes a continuous phenomenon – agents constantly analyze, decide, and act on every customer in real time. For the first time, businesses can deliver endless campaigns and personalization at 1:1 scale.”
Databricks also announced an open partner ecosystem that will import and process data across platforms and partners, including Adobe, Meta (audience and conversions API), Acxiom, Epsilon, LiveRamp, Trade Desk, Braze, Bloomreach, Iterable, Snapchat, Magnet, TransUnion, Adgra, Twilio, Twilio, Science
Native integration and reversible ETL allow CustomerLake users to connect Databricks with bidirectional pipelines across the marketing and advertising technology stack.
The CustomerLake identity solution integrates rules and agents to combine disconnected records into richer, more accurate customer profiles. A built-in identity marketplace enriches profiles with third-party identity graphs and partner data from providers such as Acxiom, Epsilon, LiveRamp, TransUnion, and Adstra.
