Contract or Code: The Role of AI Forecasting-as-a-Service in E-Commerce Platforms

This study examines suppliers' adoption decisions regarding AI-driven Forecasting-as-a-Service (FaaS) within e-commerce platform supply chains, with a particular focus on how different supply chain contracts—wholesale pricing and agency pricing—affect such decisions. Based on a game-theoretic supply chain model, the research analyzes how FaaS influences production decisions, profit distribution, and social welfare. Key findings include: First, the introduction of technology cannot offset the deep impact of supply chain contract. The agency contract consistently outperforms the wholesale contract by eliminating double marginalization, while FaaS has limited ability to alter the price–demand relationship. Second, good technology is not always necessarily worth adopting. The value of FaaS depends on the magnitude of losses caused by quantity–demand deviation; when such losses are small, the corrective benefits of FaaS may not justify its subscription cost. In addition, we identify a strategic misalignment between suppliers and platforms. However, by combining the agency contract with FaaS, both parties may achieve a win–win outcome. Finally, the impact of FaaS adoption on supply chain performance—particularly supplier profit and social welfare—exhibits a non-monotonic relationship with respect to technology development cost and subscription fees. Managers should avoid blindly pursuing “extreme optimization” of technology.