Examining customer usage intention of AI assisted chatbots in e-commerce market through customer satisfaction and trust

Artificial intelligence (AI) chatbots, powered by large data models, have become an essential tool for improving user experience and are increasingly involved in providing customer assistance and recovering from service failures. This study investigates how AI chatbots are used in the United Kingdom (UK) e-commerce sector, with a particular emphasis on how they affect user engagement and service recovery. Drawing upon Expectation Confirmation Model (ECM) and Technology Acceptance Model (TAM), a research framework is developed to investigate how anthropomorphism cues, perceived ease of use (PEOU), and problem-solving ability (PSA) affect user satisfaction, trust, and continued usage intention (CUI) of AI chatbots. The study utilizes structural equation modeling (SEM) technique to evaluate data gathered via an online questionnaire from 348 customers who had used AI chatbots for service recovery. Analyzed data using Smart PLS 3, the findings reveal that anthropomorphism cues, such as human-like behaviors and PEOU significantly increase user satisfaction and trust, which in turn increases CUI. Furthermore, problem-solving ability significantly affects satisfaction but shows an insignificant relationship with trust. Moreover, the findings also revealed that satisfaction and trust act as key mediators between these antecedents and CUI. The study enriches our theoretical knowledge of how users perceive and behave in AI-driven interactions on e-commerce platforms.