The relational power of friendly AI shopping assistants: A serial mediation model of sense of care, reliance, and stickiness intention

The rapid rise of generative AI has produced conversational AI shopping assistants that provide socially expressive, curated guidance in online retail. Yet existing research continues to focus largely on functional, service-oriented chatbots, offering limited insights into how these assistants shape consumers' relational and emotional responses. Drawing on attachment theory, this study examines how consumers' perceptions of AI shopping assistant personality—professional versus friendly—affect their stickiness behavior through perceived care and psychological reliance. Two experimental studies are conducted: Study 1 finds that friendly (vs. professional) AI assistants elicit greater stickiness intention, which is explained by a sequential pathway in which perceived care increases reliance, in turn enhancing stickiness. Study 2 demonstrates that decision difficulty moderates this process: the relational advantages of friendly AI assistants intensify with greater difficulty but diminish when decisions are easy. These findings extend attachment theory to consumer–AI interactions and show that emotional relational processes, not merely functional performance, are central to fostering sustained engagement with AI shopping assistants. For practitioners, the results highlight when and why deploying friendly AI personalities can strengthen consumers’ relational connections and continued use in online retail environments.