Autonomous Enterprise Automation using Agentic Workflows and Event-Driven Architecture for Zero-Trust Security

Abstract

Enterprise automation is evolving from rule-based orchestration toward autonomous systems capable of perceiving events, reasoning over contextual information, selecting actions, and adapting workflows with limited human intervention. This paper examines an architecture that combines agentic workflows, event-driven architecture, and Zero-Trust Security to support secure autonomous enterprise operations. The proposed approach treats security as a continuous decision process rather than a perimeter-based control. Agentic components monitor business and infrastructure events, interpret contextual signals, invoke authorized tools, and execute workflow actions while operating within explicitly defined identity, policy, and privilege boundaries. Event-driven architecture provides asynchronous communication and enables real-time propagation of security and operational events across distributed enterprise environments. Zero-Trust principles strengthen this model by requiring continuous authentication, authorization, least-privilege access, policy enforcement, and verification of agent actions. The research methodology adopts a design-oriented approach involving architectural modeling, threat analysis, workflow experimentation, and evaluation using security and automation metrics. The proposed framework can be assessed through measures including response latency, policy-violation rate, successful task-completion rate, false-positive rate, auditability, and resource utilization. The study argues that combining these technologies can create a more resilient automation environment, provided that autonomous decisions remain observable, bounded by policy, and subject to appropriate human oversight. The resulting architecture offers a foundation for secure, adaptive, and scalable enterprise automation.