AI-Driven Circular Supply Chains: Building Resilience in a BANI Landscape through Predictive Analytics

Abstract

Modern supply chains are more and more likely to be disrupted by unstable and nonlinear dynamics, which are explained in the BANI (Brittle, Anxious, Nonlinear, Incomprehensible) framework. Conventional linear supply chain models find it difficult to adjust to these circumstances, underscoring the necessity for predictive, circular, and technologically adaptable systems. This research examines the potential of AI-driven predictive analytics to bolster the resilience and efficiency of circular supply chains through enhanced forecasting accuracy, optimized resource recovery, and facilitation of proactive decision-making. A systematic review, thematic synthesis, and predictive modeling approach were utilized to evaluate the function of machine learning techniques such as Random Forest, K-Means clustering, and Neural Networks in addressing demand variability, return cycle complexity, and disruption risks. The results show big improvements, such as a 25–35% increase in the accuracy of forecasts, a 20–30% decrease in stockouts and overstocking, and a 30–40% decrease in the time it takes to get goods back. Using these ideas, a multi-layered conceptual framework was created that includes predictive analytics, circular operations, resilience feedback loops, and organizational enablers. The research finds that AI is a strategic driver for creating circular supply chain resilience in BANI settings by making it possible to see things in real time, plan for changes, and use logistics that can be reused. These results lay the groundwork for more research and real-world use in changing supply chains to be more digital.