How return policies impact the brand owner’s AI adoption strategy

Artificial intelligence (AI) is increasingly used in online retail to reduce product fit uncertainty. This paper develops a game-theoretic model with consumer learning to investigate how platform return policies shape AI adoption and market outcomes. The model assumes that a brand owner sells through an e-commerce platform. The analysis examines the brand owner’s pricing and AI adoption decisions under no-return and lenient return policies. It allows for three types of AI adoption strategies: no AI, platform-provided AI, and jointly developed AI. The results show that return policies alter the economic value of AI-generated information, thereby shaping the brand owner’s pricing and AI adoption decisions. Under a no-return policy, the belief enhancement effect of AI can strengthen the brand owner’s pricing power, while the profitability of AI adoption depends on the market fit rate. Under a lenient return policy, the screening effect of AI can reduce purchases by poorly matched consumers and the resulting return-processing losses. When the market fit rate is low, such screening can also reduce the attractiveness of a high-price strategy and lead to a lower equilibrium price. These differences reshape the brand owner’s AI adoption incentives across return policy environments and alter the relative attractiveness of AI options with different effectiveness and implementation costs. Consequently, the distribution of benefits among market participants also varies across return-policy regimes. Under a no-return policy, AI adoption may be more favorable to the brand owner and the platform, whereas under a lenient return policy, it is more likely to generate a triple-win outcome for the brand owner, the platform, and consumers.