Autonomous Audit Agents for PCI DSS 5.0: A Reinforcement Learning Approach

Maintaining continuous compliance with the Payment Card Industry Data Security Standard (PCI DSS) 5.0 remains a critical challenge due to evolving threats and system changes that lead to compliance drift between annual audits. This paper introduces autonomous audit agents leveraging reinforcement learning (RL) to address this gap. The proposed agents perform real-time inspection of control telemetry, dynamically map collected evidence to PCI DSS requirements, and autonomously generate remediation pull-requests to rectify deviations. A simulated sandbox environment, replicating multi-stakeholder payment ecosystems (acquirer, processor, and merchant systems), validates the approach, demonstrating a 92% acceleration in detecting compliance deviations and 30% reduction in audit-related costs compared to traditional methods. The results highlight the potential of RL-driven automation to enhance compliance sustainability, reduce manual intervention, and improve operational efficiency in payment card security frameworks.