Artificial Intelligence (AI)-powered chatbots have become a common interface for retail banking, handling balance enquiries, fund transfers, complaint registration and basic financial guidance without human intervention. While banks have adopted chatbots primarily to reduce service cost and improve response speed, customer willingness to rely on them for banking tasks depends on perceived usefulness, perceived risk and trust in the underlying technology. This study examines customer perception and adoption intention toward AI-powered banking chatbots using a structured questionnaire administered to 120 bank customers. Consistent with the Technology Acceptance Model, the results show that perceived usefulness has a significant positive effect on willingness to use banking chatbots, while perceived risk — particularly concerns about data privacy, transaction errors and lack of human escalation — has a significant negative effect. The study concludes that accuracy, data security and an easy pathway to human assistance are the key levers banks must strengthen to expand chatbot adoption, and offers suggestions for banks, technology vendors and regulators.