Sustained Impact of Agentic Personalisation in Marketing: A Longitudinal Case Study

A new study from researchers Olivier Jeunen, Eleanor Hanna, and Schaun Wheeler provides rare longitudinal evidence on the real-world performance of agentic AI in marketing. Published as "Sustained Impact of Agentic Personalisation in Marketing: A Longitudinal Case Study" and accepted to the UMAP '26 Industry Track, the research analyzes an 11-month deployment in a consumer application. The system used agentic infrastructure—AI that can take actions—to personalize marketing messaging at scale, moving beyond traditional static, rule-based CRM strategies.

The study design compared two distinct periods: an initial active phase where human marketers directly curated content, audiences, and strategies, followed immediately by a passive phase where AI agents operated autonomously from a fixed library of components. The key finding is that while the active human management phase generated the highest relative lift in engagement metrics, the autonomous agents successfully sustained a positive performance uplift throughout the passive period. This challenges the assumption that continuous human oversight is necessary, suggesting instead a symbiotic model.

The researchers conclude that human intervention is most valuable for strategic initialization and discovery—setting up the system and identifying effective components. Once established, autonomous AI agents can ensure the scalable retention and preservation of those performance gains over time. This has significant implications for enterprise marketing teams looking to scale personalized communication without linearly scaling human labor, pointing toward a future of hybrid human-AI workflow optimization.