DVM-HALL Model Reinvents Customer Loyalty for AI Agents

A new academic paper from researchers Sai Srikanth Madugula, Peplluis Esteva de la Rosa, and Daya Shankar proposes a framework for understanding customer loyalty when AI agents act as autonomous shoppers. The Dynamic Verifiable Multi-Agent Human Agentic Loyalty Loop (DVM-HALL) model addresses the fundamental disruption caused by agentic AI — systems that don't just recommend but independently execute purchasing decisions. Traditional loyalty models fail to account for algorithmic bounded rationality and constructed autonomy, so the authors formalize brand choice using a softmax probability formulation with five distinct components: human emotional equity, agentic machine-experience utility, calibrated trust, delegated authority, and verifiable execution. The model updates these factors recursively after each interaction, allowing trust and delegation levels to shift dynamically.

To measure alignment between human intent and agent actions, the paper introduces the Net Human-Agent Score (NHAS) — an auditable, risk-weighted metric that draws on human feedback, execution logs, benchmark comparisons, and verifiable receipts. Critically, it incorporates decentralized finance (DeFi) execution risks like gas costs, slippage, MEV exposure, and smart-contract vulnerabilities as core predictors of agentic brand preference. The authors propose a three-stage empirical validation plan: controlled shopping experiments, multi-agent market simulations, and DeFi testbeds. This framework gives brands a theoretical foundation to navigate the shift toward machine customers, where loyalty loops must account for both human and algorithmic decision-making.