What future for AI chatbots in banking: from current challenges to mirages?

Purpose Technological advancement is a crucial driver of value creation and enhanced user experience in the banking industry. This study aims to better understand the challenges of AI chatbots in banking and assess whether they herald a customer service revolution. Design/methodology/approach We analysed qualitative data collected in 2024 from three Chinese banks that implemented AI-powered chatbots. Using purposive sampling, 22 customers and managers were interviewed through semi-structured interviews guided by constructs from UTAUT2. Findings The analysis identifies five types of awe associated with chatbot use, technological, functional, conversational, service and security awe. These emotional responses are largely absent from IS literature, despite their potential to positively or negatively shape adoption. Our results suggest that while chatbots excel in routine transactions, they lack empathy and flexibility in complex cases. Functional richness and anthropomorphic features often trigger awe, which may foster engagement but also fear or discomfort. Research limitations/implications The study is exploratory and based on a limited sample. Future work should triangulate with additional qualitative evidence and employ quantitative methods to validate the role of awe in adoption models. Practical implications Banks should invest in advanced AI to improve personalization, transparency and emotional sensitivity, while maintaining hybrid systems that escalate complex issues to human representatives. Originality/value This study contributes to IS literature by integrating awe into technology adoption frameworks, offering new insights on human, AI interaction in financial services and highlighting value creation mechanisms through chatbots.