Will customers buy financial products impulsively? AI-powered chatbots’ recommendations

Purpose The purpose of this study is to examine what factors can drive financial institution (FI) customers’ impulse buying decisions in artificial intelligence (AI)–powered environments. Based on the stimulus–organism–response (S-O-R) model, this study tests the influences of perceived human-likeness and perceived interactivity on trust in AI chatbot recommendations and impulse buying, perceived warmth and perceived competence as mediators and need for human interaction as a moderator. Design/methodology/approach Sample data for this study were collected from Taiwanese customers who had received product recommendations from AI-powered chatbots provided by financial institutions in Taiwan. A final sample size of 388 usable questionnaires was obtained, and structural equation modeling was conducted to validate the research model. Findings This study showed that FI customers’ perceived human-likeness and perceived interactivity positively influenced their perceived warmth and perceived competence, which together explained their trust in AI chatbot recommendations, and in turn uplifted their impulse buying. Besides, this study showed that FI customers’ need for human interaction significantly moderated some path relationships in the research model. Originality/value Based on the SOR framework, this original study contributes to the literature for understanding the formation of FI customers’ trust in AI chatbot recommendations and impulse buying by integrating the computers-are-social-actors paradigm, stereotype content model and elaboration likelihood model.