The most intellectually provocative story of the week came from Minerva, an AI startup that has built what AdExchanger described as an 'agentic CDP' — a platform that orchestrates customer data workflows autonomously — while explicitly rejecting the CDP label. The framing is strategic: CDP as a category carries baggage (failed implementations, data silos, integration debt), and Minerva is betting that 'agentic' is a more compelling and accurate descriptor for what AI-native data platforms actually do. This mirrors a broader industry pattern where AI-first companies are redefining legacy categories rather than competing within them.
Complementing this, the week's conversation around Marketing Mix Modeling highlighted AI's limits as much as its promise. AdExchanger's analysis argued that while MMM is getting smarter — incorporating more granular signals, faster refresh cycles, and machine learning — it still cannot replace the full measurement stack on its own. The piece is a useful corrective to AI hype: augmentation, not replacement, remains the operative model. PurePlay AI's Eric Steigelfest made a similar point in a Beet.TV appearance, arguing that 'spray and pray' campaign logic needs to give way to precision targeting — but that precision requires better data infrastructure, not just better algorithms.