Exploring the User Experience and Cognitive Absorption with Conversational Agents: The Effects of Conversational Norms in Human-AI Interactions Through a Mixed-Methods Study

In this work we investigated the interplay between conversational user experience (UX) dimensions, conversational agent (CA) capabilities, and cognitive absorption (CGA) in human-AI interactions. Leveraging the Gricean maxims (quality, quantity, relevance and manner), its AI-specific extensions (benevolence, transparency and priority), our previously developed measurement scale called CASUX (Conversational Agent Scale for User Experience), and the Cognitive Absorption Theory; this research explores how different conversational capability agents influence users’ CGA. We undertook a cross-sectional comparative study, and analyzed using a mixed-methods approach by combining Partial-Least-Squares Structural Equation Modeling and Fuzzy-set Qualitative Comparative Analysis. Two custom-made CAs were developed using the Alexa skills kit, each representing distinct conversational capabilities (baseline vs. customized). Our findings reveal significant differences between CA types with proficiency emerging as the critical functional driver of CGA for customized CAs, whereas baseline CAs rely on social UX dimensions such as humanness, anthropomorphism, etiquette & mannerism, and personality.