Artificial Intelligence-Driven Forecasting Practices and Supply Chain Performance: Toward a Conceptual Framework

Abstract

In an economic environment characterized by increasing demand volatility, conventional forecasting methods are revealing their limitations, prompting growing interest in demand sensing. While artificial intelligence (AI) is widely regarded as a catalyst for this transition, the precise mechanisms through which it transforms practices and impacts supply chain performance remain inadequately elucidated in the academic literature. This study aims to address this gap by introducing the "AI-Demand Sensing-Performance" (ADPsc) conceptual model, which delineates the relationships between AI capabilities, the implementation of demand sensing, and the enhancement of supply chain performance. Grounded in a deductive methodology rooted in contingency theory, this research constructs a theoretical framework articulated around three interdependent pillars: (1) real-time, multi-source data integration, (2) processing through AI algorithms, and (3) operational activation, all driven by a continuous learning loop. The ADPsc model highlights the moderating role of AI in the impact of demand sensing on performance. It demonstrates that AI serves not merely as an optimization tool but as the foundation for a continuous adaptive capability, thereby enabling unprecedented responsiveness to demand fluctuations.