Artificial intelligence is the technology that systematically simulates human thinking and decision to accomplish some particular tasks more efficiently based on big data and machine learning. Benefiting from the advancement of the Internet of Things（IoT）and cloud computing in Internet era, artificial intelligence is developing rapidly to satisfy commercial requirements, which also draws academic attention. Even though it is commonly believed that artificial intelligence has reshaped the fit of business model, the mechanism remains obscure. Therefore, grounded in the specific context of artificial intelligence commercialization in the personalized recommendation of retail e-commerce platforms, from the perspective of customer-product-fit, and based on the theoretical foundation of service dominant logic, this study sets Pinduoduo, a new e-commerce platform, as the research object, and defines the business model as an organic whole consisted of “customers”, “products” and “store”（customer-product-fit）according to business model canvas and the essence of retailing, then a case study is conducted to investigate how artificial intelligence reshapes the fit of business model. The findings demonstrate that with the transformation from goods dominant logic to service dominant logic, artificial intelligence reshapes “customers” by making them central, social and contextual to describe the characteristics of customers, and reshapes “products” by making them online, precise and emotional to enrich the attributes of products. More specifically, recommendation algorithms based on content and collaborative filtering are adopted to precisely match the characteristics of customers and the attributes of products. For “customers”, centralization is the foundation of socialization, socialization is the support of contextualization, and contextualization further enhances centralization in turn; for “products”, online is the prerequisite of precision, precision is the guarantee of emotionalization, and emotionalization further improves online in turn, thus turning the path of “store”（customer-product-fit）from “customers searching for products” to “products searching for customers”. Furthermore, the features behind “store”（customer-product-fit）are revealed by the synergic reshape between “customers” and “products”, namely, the way they fit turns from “products of an attribute flock together” to “customers of an attribute flock together”, and the extent they fit turns from “thousand customers one face” to “one customer thousand faces”.
This study contributes to the extant literature in three aspects: First, it enriches relevant studies on how artificial intelligence reshapes business model at the component level through modular reconfiguration according to business model canvas and industrial characteristics. Second, it refines the massive studies on the fit of business model by examining its forming path, pattern and extent from the perspective of customer-product-fit. Third, the mechanism behind how artificial intelligence reshapes the fit of business model is clarified through the theoretical foundation of service dominant logic under the specific context of artificial intelligence commercialization. In conclusion, drawing on modular reconfiguration of business model, research perspective of customer-product-fit and theoretical foundation of service dominant logic, this study not only opens the “black box” of how artificial intelligence reshapes the fit of business model, but also sheds light on the innovation of retail e-commerce platforms’ business models empowered by artificial intelligence both theoretically and practically.