Chatbots have been in hotels and travel platforms for nearly a decade, long enough for a meaningful body of research to accumulate about what they actually do for businesses and guests. This paper takes stock of that research through bibliometric analysis — systematically mapping the key themes, most-cited findings, and emerging directions in chatbot research across hospitality and tourism.
The core finding that will resonate with operators: chatbot adoption in hospitality is not primarily a technology question anymore. It's a service design and trust question. The research consistently shows that guests are willing to use chatbots — even prefer them in certain situations — but their satisfaction depends heavily on factors that technology vendors rarely lead with: perceived competence (does the bot actually understand what I'm asking?), transparency (do I know I'm talking to a bot?), and recovery capability (when the bot fails, what happens next?).
Three clusters dominate the research landscape, representing the areas where the evidence is strongest. First, service quality: chatbots that are well-designed and clearly scoped (handling booking queries, FAQ responses, and simple service requests) reliably deliver high satisfaction scores. Chatbots that are deployed too broadly, with poorly defined boundaries, generate frustration that can damage the overall service perception. The clearest practical guidance here is to define the chatbot's scope narrowly and communicate it clearly to guests.
Second, customer engagement: chatbots have proven particularly effective at the top and bottom of the customer journey. Pre-arrival, they handle high-volume routine questions that would otherwise consume front desk time. Post-stay, they can solicit feedback, offer loyalty program information, and handle routine post-stay service issues. In the middle of the stay, human interaction generally still outperforms chatbot interaction for anything requiring nuance or emotional intelligence.
Third, trust: this is where the research identifies the biggest risk for operators. Guests who feel deceived — who discover mid-conversation that they were talking to a bot without being told — report significantly lower trust in the brand, not just the tool. The paper strongly recommends transparency-by-default: make the bot identity clear upfront, and ensure guests can reach a human easily when they want to.
The bibliometric analysis also surfaces a growing research interest in upselling and revenue generation through chatbot interactions. Early evidence suggests that well-timed, contextually appropriate upgrade offers delivered through chatbots can generate incremental revenue without the awkwardness that human staff often feel around upselling. This is an area worth watching and piloting.
For technology buyers, the paper's synthesis is useful for evaluating vendors: the research distinguishes between basic keyword-matching chatbots (low cost, limited capability, lower guest satisfaction) and conversational AI systems based on language understanding (higher cost, better comprehension, significantly better satisfaction outcomes). The gap between these two categories has narrowed on price but widened on performance, making the upgrade case increasingly compelling for high-volume properties.