Effects of perceived media richness on retail chatbot user experience

Purpose This study investigates how chatbot media richness shapes shoppers' perceptions of social presence, emotions, trust and their willingness to recommend e-retailers’ chatbots. Design/methodology/approach Data were collected from a random sample of 460 pre-recruited online shoppers in South Africa. A structured online survey was deployed. To test the research model, ADANCO was applied to run confirmatory composite analysis using modern partial least squares structural equation modelling (PLS-SEM). Findings The results reveal perceived media richness positively and significantly influences perceived social presence, emotions and trust in chatbots. Our study concedes on the effects of social presence on human emotions, yet challenges how the former inform users' trust perceptions, suggesting other possible mechanisms for trust formation in chatbot contexts. Trust and emotions are significant predictors of the willingness to recommend chatbots. Research limitations/implications A cross-sectional, single-country study via self-reports was conducted with online shoppers, which could limit the findings' generalisability. Future research could validate the findings through longitudinal designs and cross-cultural replications. Practical implications Our study provides significant insights into the complex dynamics of human–chatbot interactions, yielding unique psychological and behavioural outcomes. The study offers practical implications for chatbot design, fostering emotional connections and trust-yielding social presence and informing chatbot usage recommendations. Social implications Nevertheless, chatbot-mediated environments may produce complex, and sometimes contradictory societal implications, including privacy concerns, emotional overreliance and the displacement of some human-human interactions. Originality/value This study advances theory and practice by clarifying how media richness, emotional engagement, social presence, trust, shoppers and user outcomes interact within the context of e-retailers’ chatbots.