Conversational AI in hospitality and tourism: a bibliometric–systematic review

This study examines how conversational artificial intelligence (chatbots, virtual assistants, and large language model–based agents) shapes the customer journeys of travellers, guests, and visitors in hospitality and tourism, including hotel, online travel agency, and airline services. We consolidate dispersed findings into an integrated account that links conversational design levers, psychological mechanisms, outcomes, and contextual boundaries. Using a protocolised bibliometric–systematic review of Scopus (search window 2010–2025; final n = 71 journal articles), we apply transparent eligibility criteria, quality appraisal, structured coding, and bibliometric mapping across booking, concierge/guest service, recommendation/personalisation, and service recovery touchpoints. Evidence indicates that response speed, message concision, empathic style, anthropomorphism, disclosure/explanation practices, visual/emoji enrichment, privacy-by-design features, and hybrid human handoff influence trust, perceived competence, reduced expectancy violation, perceived control, inspiration, and rapport. These mechanisms are associated with adoption and continuance, purchase and recommendation, satisfaction and loyalty, recovery effectiveness, and willingness to disclose data, with effects moderated by task complexity, failure severity, user traits, setting, and culture/digital literacy. From a practical perspective, this synthesis also provides guidance for hospitality and tourism practitioners on aligning large language model (LLM) capabilities with customer-journey touchpoints, calibrating response time and message concision to task and social context, implementing privacy-preserving personalisation, and maintaining human involvement in high-stakes or emotionally sensitive interactions.