AI-Augmented Vendor Risk Management: Continuous Assessment of Third-Party Security across Global Retail Supply Chains

The global retail ecosystem is increasingly relying on multi-tier suppliers, including third, fourth, and fifth-party vendors, which help provide logistics, point-of-sale platforms, ecommerce platforms, and digital functions. With these dependencies, their opposition to vendors grows as the most effective way to initiate large-scale cyber intrusions, as seen with the SolarWinds, MOVE it, and Okta breaches. Traditional vendor risk management operates on a basis that utilizes the static and questionnaire approach to assessments that cannot be used to determine real-time security posture or the dynamism of threats. This study suggests an AI-enhanced Vendor Risk Management (VRM) system combining behaviour analysis based on machine learning, dependency mapping by graph neural networks, and ongoing compliance, vulnerability, and exposure to the supply-chain analysis. The model contains onthe-fly risk scoring, early anomaly detection, and predictive threat modelling in a multi-tier vendor environment. Its proposed architecture can be used as critical infrastructure in retail supply chains, allowing for the continuous enforcement of contracts, prediction of fraud risks, and automatic execution of remediation processes. This study illustrates that AI-based VRM can provide resilience, minimize supply-chain vulnerabilities to attacks, and provide an operational base of zero trust in distributed retail ecosystems through empirical performance assessment and structural assessment of architectural systems. The results indicate that realtime and AI-based risk evaluation is required to safeguard globally connected supply chains and predict antagonistic actions in the future.