Zero Trust Architecture with MLDriven Behavioral Analytics for Healthcare ERP on Microsoft Azure

Abstract

Healthcare ERP systems on Microsoft Azure require access controls that can evaluate not only user identity, but also session context, privilege behavior, transaction activity, and network movement. This article presents a Zero Trust architecture enhanced by ML-driven behavioral analytics for adaptive access enforcement in healthcare ERP workloads. The framework integrates Azure AD sign-ins, MFA status, conditional access decisions, RBAC assignments, privileged role activation, device compliance, ERP transaction logs, NSG flow records, private endpoint activity, Sentinel alerts, and runtime telemetry. ML models generate user behavior risk scores, privilege deviation scores, session trust probabilities, ERP transaction anomaly scores, and network behavior threat scores to support continuous verification. Results show that identity trust accuracy, privilege deviation detection, and ERP transaction anomaly recognition improve across Zero Trust verification cycles, while healthcare ERP access scenarios reveal distinct variations in user behavior risk, session trust, and network threat exposure. The study concludes that Zero Trust becomes more effective for healthcare ERP security when Azure-native controls are combined with behavioral intelligence and ERP-specific transaction monitoring.