Abstract
Intelligent Virtual Assistants (IVAs) powered by Artificial Intelligence (AI) are increasingly embedded in tourism and travel services, supporting tasks from itinerary planning and booking to post-purchase assistance. Despite their growing adoption, research has yet to fully explain how perceived IVA intelligence drives continuance usage across customer journey stages and service contexts. Existing studies treat IVAs largely as generic information systems, emphasizing utilitarian outcomes such as efficiency and perceived usefulness, while neglecting the multidimensional nature of intelligence. Social, conversational, and emotional intelligence are rarely distinguished, leaving limited understanding of how these dimensions jointly and differentially influence user satisfaction and continued engagement. Furthermore, prior work rarely considers the dynamic nature of tourism interactions across pre-purchase, purchase, and post-purchase stages. User expectations, priorities, and emotional responses shift across these stages, influencing evaluations of IVA performance. Similarly, the hedonic versus utilitarian service distinction remains underexplored. Emotionally expressive or socially engaging IVAs may enhance engagement in hedonic contexts (e.g., luxury travel planning) but be perceived as inefficient in utilitarian contexts (e.g., budget airline bookings). Current models do not account for such contextual and stage-specific contingencies, limiting both theoretical insight and managerial guidance. This development paper proposes a stage-sensitive, context-contingent framework that integrates multidimensional IVA intelligence, socio-psychological mechanisms, specifically user pleasure and task–technology fit, and anthropomorphism as a moderator of satisfaction. The framework will be empirically examined through a three-stage, cross-national design with data from the United Kingdom and the United States (n = 900). Multi-group Structural Equation Modelling (SEM) will assess variation across customer journey stages, service contexts, and national settings, as well as the moderating influence of anthropomorphism. By capturing stage-specific, context-dependent, and socio-psychological mechanisms, the study advances theory by refining the conceptualization of IVA intelligence, elucidating how satisfaction and continuance are jointly shaped by intelligence, pleasure, and task–technology fit, and demonstrating the importance of cross-contextual generalizability. For managers, the findings will provide actionable guidance for designing IVAs that are not only functionally effective but socially and emotionally intelligent, tailored to both hedonic and utilitarian service contexts, and responsive across the entire customer journey, thereby fostering sustained user engagement and loyalty.