AI startup Minerva built an agentic CDP, but don't call it that

Minerva, which raised US$20 million to build AI agents for marketers on an identity graph of 270 million consumers, is betting that the customer data platform gets absorbed into an agent layer rather than remaining a dashboard. CEO Jackson Engles, a former Lazard banker, is candid that the data is not the moat: "Yes, sure, we have data that can help create better outcomes... But is it a long-term, durable moat? Probably not." The founders came from finance, Engles from Lazard, Daniel Saedi from Bridgewater and Mathew Joseph from Citadel, and their pitch is that a specialized, opinionated data model beats general-purpose AI for marketing. Minerva's agentic data engineer maps a brand's CRM and warehouse, then supplements it with a B2B2C identity graph linking work and personal identities; client Ramp used that link to extend a direct-mail campaign and lifted its marketing-qualified lead rate about 80% in three months, while Wander's lookalikes drove a 3.4x higher return on ad spend than Meta's broad targeting.

The crux

The CDP is being reimagined as an agentic middle layer where identity resolution and automation, not a query dashboard, deliver the outcome, in what Engles calls the post-CDP era.

Sell-side read

As brands consolidate on agentic CDPs, publisher first-party data becomes more valuable as identity-graph enrichment, but the margin on that enrichment accrues to the vendor, not the publisher, unless publishers price their data access accordingly.

Watch

Whether Minerva's second agent, an agentic data scientist for predictive modelling, ships, and how incumbents like Salesforce respond.