CustomerLake: The Agentic CDP Built for What Marketing Has Become

For most of the past two decades, enterprise marketing ran on a relatively stable loop: collect customer data, build a segment, craft a message, launch a campaign, measure results, and repeat. The tools that powered this loop — CRMs, CDPs, marketing automation platforms, data warehouses — were each designed for a specific layer of that process. Individually, they were useful. Together, they were slow, fragmented, and increasingly hard to reconcile.

It Started with CRM

Customer Relationship Management systems emerged to give businesses a structured way to track individual customer interactions — sales history, contact records, support tickets, and deal stages. CRM answered a simple and valuable question: who is this person, and what have we done with them? It was a contact-centric model built for sales teams managing human relationships one at a time.

But as digital channels multiplied and marketing grew more data-intensive, CRM’s limitations became apparent. It captured structured relational data but couldn’t handle the volume, velocity, or variety of behavioral signals generated by a web-scale customer base. It also couldn’t unify data across disconnected touchpoints — a purchase in one system, a support call in another, a web session in a third. CRM was a ledger, not a living 360-degree view of the customer.

Then, the CDP

That gap is what gave rise to the Customer Data Platform. The CDP evolved directly out of CRM’s limitations. While CRM organized interactions by account or contact, the CDP was designed to unify customer data from every source into a single, persistent profile — the Customer 360. It pulled behavioral data, transaction records, engagement history, and third-party enrichment into one place, then activated that unified view across marketing channels. The CDP answered a more sophisticated question: who is this person across all of the ways they interact with us, and how do we reach them?

For a decade, that was enough. It isn’t anymore.

The reason isn’t that CDPs became poorly built — it’s that the problem fundamentally changed. Consumers are beginning to deploy their own agents to research, compare, and transact on their behalf. A buying decision that once took days or weeks now collapses into milliseconds. Batch-based campaigns, static segments, and manually assembled customer journeys are invisible to this new buying behavior by the time they respond.

Existing CDPs also share an architectural weakness that this new era exposes completely: they sit outside the enterprise data platform. Every time a marketing team needs fresh customer context, data has to be moved, copied, and reconciled across systems — creating latency, governance complexity, and the constant risk that the profile the CDP is working from is already out of date.

Marketing has always evolved its infrastructure when buyer behavior outpaced existing tools. CRM was the answer when the problem was organizing human relationships at scale. CDP was the answer when the problem was unifying fragmented digital identities. The agentic era requires something that neither was built to deliver.

CustomerLake

Databricks CustomerLake is a native Agentic Customer Data Platform embedded directly in the Databricks Lakehouse — bringing Customer 360, identity resolution, audience building, campaign automation, activation, and personalization into the same governed environment where customer data, AI models, and business logic already live.

CustomerLake was announced at the 2026 Data + AI Summit and is currently available in Private Preview.

A Single Foundation for Marketing, Data, and AI

The core problem CustomerLake addresses isn’t a feature gap — it’s an architectural one. When the CDP sits outside the data platform, marketing and data teams are perpetually out of sync. Data engineers spend cycles on ad hoc exports. Marketers wait on request queues. Sensitive customer data gets copied into systems that operate under different governance rules. The customer profile that actually powers a campaign is always, to some degree, a snapshot of the past.

CustomerLake eliminates that gap by living inside Databricks. Customer context — spanning transactions, behavioral signals, product usage, support history, loyalty data, and third-party enrichment — is unified, governed, and accessible in one place. AI models and agents operate on that same foundation. Marketing doesn’t need to wait for data engineering to prepare a dataset before it can run a campaign. And because everything runs under Unity Catalog, the same entitlements and security boundaries that govern every other operation in the enterprise apply equally to every agent action in CustomerLake.

The result is a new operating model for marketing — one where agents continuously analyze customer signals, determine the next-best action, and execute across channels in real time, while human marketers define the strategy, goals, and guardrails. Campaigns stop being discrete, manually rebuilt events and become continuous, adaptive systems.

With CustomerLake, we’re replacing legacy software with an open, Agentic CDP built directly on the Lakehouse. When customer data, AI models, and agents live in one governed platform, marketing stops being a series of campaigns and becomes a continuous loop — agents that constantly analyze, decide, and act on every customer in real time. For the first time, enterprises can deliver true 1:1 experiences at an infinite scale.

Ali Ghodsi, Co-Founder and CEO, Databricks

What CustomerLake Delivers

CustomerLake capabilities span the full range of what enterprise marketing teams need, from data unification through activation and personalization. Here is what the platform provides:

• Agentic Identity Resolution (AIR): A new approach to identity that combines deterministic, probabilistic, and agentic workflows to unify disconnected customer records into accurate, continuously improving profiles. Teams can bring existing identity rules and third-party enrichment partners, while Profile Agents surface edge cases and maintain quality over time.
• Campaign Agents: AI agents that move marketers from static, one-off campaigns to always-on engagement. Campaign Agents draw on governed customer context in Databricks to build audiences, recommend next-best actions, activate across channels, and continuously optimize toward defined business goals. Marketers set the strategy and guardrails; agents handle the execution at scale.
• Customer 360 Profiles: Business-ready unified customer profiles built and maintained directly inside Databricks, drawing on the full breadth of enterprise data — transactions, behavior, product usage, support records, loyalty, commerce signals, and third-party enrichment — without duplicating or moving sensitive data.
• Embedded Governance via Unity Catalog: Because CustomerLake runs within Databricks and is governed by Unity Catalog, every agent action operates under the same data entitlements and security boundaries as other enterprise operations. No parallel ruleset for the CDP, no separate governance system to maintain.
• Infinity Campaigns: A new engagement model that replaces the manual campaign cycle — define, segment, build, launch, measure, repeat — with autonomous, continuously adapting engagement loops. An Infinity Campaign is always on, reacts in real time, and uses LLMs and agents to personalize at the individual level rather than the segment level.
• Lakehouse Federation: Teams can access trusted customer data wherever it resides — Databricks, Snowflake, Google BigQuery, cloud object storage, operational databases, or other enterprise systems — without creating new silos or requiring data migration.
• Open Partner Ecosystem: CustomerLake launches with native integrations across identity, activation, measurement, and customer experience, including Adobe, Meta (Audience and Conversions API), Braze, Acxiom, Epsilon, The Trade Desk, LiveRamp, Iterable, Bloomreach, Snapchat, Magnite, TransUnion, Adstra, Twilio, Integral Ad Science (IAS), and Unity. Services partners include Accenture, Deloitte, Lovelytics, Slalom, and Stitch.
• Profile Agents: AI agents that help marketing and data teams turn raw customer data into business-ready Customer 360 profiles directly in Databricks. Profile Agents handle data preparation, surface quality issues, and support third-party data enrichment to continuously improve profile accuracy.

CustomerLake also introduces a value-aligned consumption model, a more flexible and cost-effective alternative to traditional software licensing — a practical consideration for enterprise brands looking to consolidate their martech stack.

Taken together, CustomerLake provides enterprise marketing teams with a single environment in which customer intelligence, AI-driven decision-making, and cross-channel activation operate without handoffs, latency, or governance gaps. The platform is built for organizations that need to engage customers — and their agents — at the speed and scale the current era demands.

At HP, we believe the future of AI-driven customer engagement depends on moving beyond fragmented customer data toward governed customer context. Databricks CustomerLake brings that vision to life, enabling HP to build customer intelligence, personalization, and activation on the data foundation we already trust, rather than creating another place where data must be copied, reconciled, and secured. With CustomerLake, marketing can move faster, operate smarter, and transform with AI using the same trusted customer context as finance, product, sales, and operations.

Kumar Ram, Global Head of Marketing Technology and AI Enablement, HP

Marketing Built for What Comes Next

The evolution from CRM to CDP to Agentic CDP isn’t a story of incremental improvement — it’s a story of infrastructure catching up to how buyers actually behave. CRM organized relationships. CDP unified profiles. CustomerLake closes the final gap: it brings the governed customer context, AI models, and execution capability into a single place, so marketing can operate at the speed of a buyer who no longer waits.

For enterprises ready to move past disconnected systems, manual campaign cycles, and customer profiles that are always slightly out of date, CustomerLake represents the architecture that the next decade of marketing actually requires.

Get Started with Databricks CustomerLake

Frequently Asked Questions