Artificial Intelligence in Tourism Through Chatbot Support in the Booking Process—An Experimental Investigation

AI-controlled chatbots have been used in travel services for some time and range from simple hotel reservations to personalized travel recommendations. However, the acceptance of chatbots compared to human interlocutors has not yet been extensively studied experimentally in the tourism context. In this experimental, randomized, vignette-based, preregistered 2 (agent: AI chatbot/human counterpart) × 3 (situation: positive/neutral/negative) between-subjects design, we hypothesized that booking intention is reduced in chatbots compared to human agents and in situations where the booking can only be made under more negative than the original conditions. Additionally, we hypothesized an interaction effect between agent and situation, presuming that the decrease in booking intention in negative situations would be less strong for human agents than for chatbots. Structural equation modelling of the data indicates strong support for the Technology Acceptance Model in the booking context. As presumed, the booking intention was lower in the negative situation and borderline lower for the chatbot. The interaction effect was shown descriptively in the data. Chatbots are recognized during the booking process and less accepted to support bookings than their human counterparts. Therefore, managers should design chatbots as human-like as possible to avoid losing sales when outsourcing customer contact activities to AI technologies.