Chatbots on the Frontline: The Imperative Shift From a “One-Size-Fits-All” Strategy Through Conversational Cues and Dialogue Designs

Abstract

The lack of transparency in AI-related technology poses challenges in identifying elements that influence conversation fluency with chatbot. Drawing from media richness, task-technology fit, and flow theories, we propose an integrated framework to investigate how chatbots’ humanoid characteristics affect users’ process fluency. Furthermore, we explore boundary conditions of dialogue characteristics, including conversation types (topic-related vs. task-related) and interaction mechanisms (free-text vs. button-based) that amplify or disrupt such flow-like experience in conversation. Two separate scenario-based experimental studies were conducted to explore two chatbot humanoid characteristics, human-like cues (Study 1) and tailored responses (Study 2). Results suggest that a match between chatbot’s humanoid and dialogue characteristics can increase fluency in comprehending the message, enhancing customer satisfaction and usage intention. Specifically, chatbots with humanoid conversational cues promote more flow-like messages in topic-related conversation or free-text interaction. The results highlight the significance of process fluency leading to more favorable outcomes in human–chatbot interactions.