AI chatbot quality in customer service: Extending measurement models with conversational capability

The development of artificial intelligence (AI) is driving the use of chatbots to improve customer service. However, measuring the quality of chatbots remains a crucial issue despite the importance of customer satisfaction. This research aims to develop a model for measuring the quality of AI chatbot customer service by integrating technical dimensions, conversational capability, and anthropomorphism to align with advancements in modern AI chatbots. This research adopts established models such as SERVQUAL, DeLone and McLean, HOT fit, and AICSQ. The quantitative approach uses Partial Least Squares Structural Equation Modeling (PLS-SEM). Data was collected in two stages: a pretest with 83 respondents and a nomological test with 213 respondents using the Telkomsel Veronika chatbot. The test results show that only the information quality and system quality variables significantly directly influence satisfaction, and only the conversational capability variable directly influences satisfaction and trust, which strengthens the intention to continue using the chatbot. This research contributes to the development of a chatbot evaluation model by highlighting the importance of technical and conversational capability dimensions in shaping user experience. Further research could expand the context to various service sectors, as well as consider local cultural and linguistic factors.