Key points are not available for this paper at this time.
Purpose This study investigates how the design characteristics of artificial intelligence (AI)-powered chatbots influence user performance in mobile banking, focusing on the mediating role of synchronous service quality. Understanding how consumers collaborate with AI chatbots in real time is increasingly critical as service interactions become more automated. Design/methodology/approach Data were collected from 263 mobile banking users with prior experience using AI-powered chatbots; 230 valid responses were retained. Partial least squares structural equation modeling (PLS-SEM) was employed to examine both the antecedents and outcomes of synchronous service quality in AI–human collaboration. Findings The results show that all three chatbot characteristics – anthropomorphic design, perceived intelligence and perceived safety – significantly shape users’ perceptions of synchronous service quality, which in turn affects task-oriented performance outcomes in terms of both emotional and cognitive responses. Perceived synchronicity emerges as a key mechanism enabling effective AI–human collaboration. Research limitations/implications The study was conducted within a single digital service context using a cross-sectional design. Future research should examine other service domains and cultural contexts, and apply longitudinal methods to improve generalizability. Practical implications Emphasizing human-likeness, cognitive responsiveness and safety in chatbot design can enhance synchronous communication, thereby improving user engagement and service efficiency in mobile banking. Originality/value This study makes a novel theoretical contribution by integrating Media Synchronicity Theory with the Computers Are Social Actors paradigm to explain how chatbot design drives real-time interaction quality. It shifts the analytical focus from user attitudes and adoption to the dynamics of interaction effectiveness – addressing a critical and underexplored dimension of intelligent service systems.
Huang et al. (Thu,) studied this question.