Author: Sreeharsha Alagani: Senior Client Partner
In my conversations with marketing and data leaders over the past year, one question keeps surfacing in different forms: Why does acting on customer data still feel slow when we have more of it than ever? The answer is rarely a missing tool. It is that the customer data sits in one place, the models that could use it sit in another, and the systems that actually reach the customer sit somewhere else again. Every campaign becomes a small integration project. At the Databricks Data and AI Summit in June, Databricks put forward a direct answer to that gap with CustomerLake, and it is worth marketing leaders understanding what it changes, even though it is still early.
Customers now treat personalization as the default setting, not a pleasant surprise. McKinsey finds that 71 percent of consumers expect personalized interactions and 76 percent get frustrated when they do not receive them. The reward for meeting that expectation is measurable, not soft: McKinsey estimates personalization can lift revenue by 5 to 15 percent, reduce customer acquisition costs by as much as 50 percent, and improve marketing return on investment (ROI) by 10 to 30 percent. Yet most teams I work with are not held back by ambition or by a lack of data. They are held back by how scattered that data is and how long it takes to turn it into something a campaign can act on. The real question sitting underneath the tooling question is about the foundation.
CustomerLake is Databricks' agentic CDP, built natively inside the lakehouse and governed by Unity Catalog. In plain terms, that means the customer data, the artificial intelligence (AI) models that act on it, and the agents that run campaigns all live in one governed place, rather than in separate systems that need constant reconciliation. It brings identity resolution, a unified customer view, audience building, campaign automation, and activation together, with Profile Agents assembling customer profiles and Campaign Agents running always-on engagement across channels.
The shift it points to is the part that matters for strategy. Personalization stops being a plan, build, ship, and measure sequence that takes weeks, and becomes a continuous loop that senses a signal and acts on it in near real time. That is a genuinely different operating rhythm for a marketing team. I want to be honest about timing, though: CustomerLake is in private preview today, so I would treat it as a direction to prepare for rather than a finished product to deploy this quarter. The strategic signal is what I would act on now.
Here is the point I am most direct about with clients. An agentic CDP is only as good as the trusted, unified customer view beneath it. Agents that act continuously on fragmented or poorly governed data do not repair that fragmentation; They act on it faster and at far greater scale. Speed applied to a shaky foundation simply produces the wrong thing sooner.
The organizations that will get the most from a platform like CustomerLake are the ones that have already done the quieter work: Resolving customer identity across sources, agreeing on the definitions and audiences everyone trusts, and putting governance in place so the right data reaches the right user safely. None of that is glamorous, and none of it depends on any single vendor's roadmap. It is also precisely what separates real-time personalization from real-time noise, and it is the difference between an agent that earns trust and one that quietly erodes it.
This is the work my team spends its time on. As a Databricks Silver Partner, KPI Partners builds governed, unified customer foundations on the Databricks Lakehouse, using Unity Catalog for governance and lineage and our enterprise analytics on Databricks to stand up a trusted customer view in weeks rather than quarters. Our agentic AI practice then helps marketing and data teams put that foundation to work responsibly, with the guardrails a regulated enterprise need.
I have seen what becomes possible when the foundation is right. For one Fortune 100 healthcare leader, we helped cut marketing campaign cycle time by 95 percent on a governed platform, compressing work that used to take weeks into a fraction of the time. That is exactly the kind of speed an agentic CDP is designed to sustain, and it is only reliable when the underlying data is trusted. The KPI Partners and Databricks partnership is where that combination of governed data and agentic delivery comes together, and it is the lens through which I would evaluate CustomerLake when it is ready.
You do not need to wait for CustomerLake to reach general availability to prepare for it, and the preparation pays off on its own. I would start in four places:
Every one of these has value today, and each compounds the moment an agentic CDP is in your hands.
The standalone CDP defined the last decade of marketing technology. The next decade looks different: The CDP comes home to the platform where your customer data and your AI already live, and personalization becomes something the business does continuously rather than campaign by campaign. CustomerLake is an early and clear signal of that shift. The leaders who prepare the foundation now, a trusted and governed customer view, are the ones who will move first when the technology is ready, and they will not be scrambling to unify their data while their competitors are already acting on theirs.
Related reading: How KPI Partners cut marketing campaign cycle time by 95 percent for a Fortune 100 healthcare leader
Ready to Transform Your Data Strategy? Talk to our experts and discover how KPI Partners can accelerate your data and analytics initiatives.
Talk to Our Experts | Book a demo