Evaluating customer engagement with AI-driven chatbot affordances in e-commerce: A multi-method analysis

With technological advancements worldwide, developing effective and efficient chatbots for e-commerce platforms is essential. This study focused on how e-commerce platforms can develop well-functioning AI-driven chatbots using three different affordances: information association, visibility, and interactivity, based on affordance theory. This study also examines the mediating effects of customer satisfaction and trust in the relationship between the three AI-driven chatbot affordances and customer engagement. Further, this study analyzed how these variables are important, their performance, and their necessity in the context of AI-driven chatbots. Using partial least squares structural equation modeling (PLS-SEM), importance-performance map analysis (IPMA), and necessary condition analysis (NCA), the results show that customer satisfaction and trust partially mediate these relationships. According to the results, customer satisfaction and trust act as partial mediators. Moreover, all the variables were found to have high performance, but only visibility and interactivity were highly important. Regarding the necessary condition analysis, interactivity was unnecessary to build customer trust. However, all variables were found to be statistically significant necessary conditions for customer engagement, with small-to-medium to large necessity effect sizes. Conducted in Sri Lanka, this study adds valuable insights into the underexplored topic of AI-driven chatbot affordances and their impact on customer engagement, offering refined insights into the mediating mechanism of customer satisfaction and trust.