Generative AI arrived in hospitality without waiting for the industry to be ready. This paper, written by some of the field's most prominent researchers, provides an early but rigorous assessment of where generative AI tools like ChatGPT actually fit in hospitality and tourism operations — and, critically, what the unresolved questions are that operators need to think through before scaling up their use.
The paper maps the opportunity across four functional areas. In guest communications, generative AI can handle first-response customer service at scale and in multiple languages — a capability that is genuinely transformative for properties that serve international guests without multilingual staff. The technology can now maintain a coherent, contextually appropriate conversation across multiple turns, handle routine booking queries and requests, and escalate gracefully to human staff when the conversation requires judgment or empathy beyond its capability.
In content creation, the productivity gains are measurable and available now. Hotel marketing teams using generative AI to produce first drafts of property descriptions, email campaigns, social posts, and review responses report significant time savings — with quality outputs that require editing rather than creation from scratch. The paper treats this as a near-certain short-term win for most operators, with the caveat that human editorial oversight remains essential and should be built into workflow design, not treated as optional.
In service design, the more interesting and less-developed opportunity lies in using generative AI to help staff rather than replace them — providing real-time information retrieval, guest history summaries, and decision support during service interactions. A front desk agent handling a complex guest complaint, for instance, can benefit from an AI tool that instantly surfaces relevant policy information, the guest's previous stays, and suggested resolution options, allowing the human to focus on the interpersonal dimension of the interaction.
In operational planning, the paper identifies revenue management, demand forecasting, and scheduling as areas where generative AI capabilities are being integrated into existing tools — with early evidence of performance improvement but not yet the rigorous comparative studies that would allow confident claims.
The second half of the paper, focused on research gaps, is where it becomes most valuable for senior decision-makers. Three issues stand out. First, the regulatory environment around AI in customer service is evolving rapidly in Europe and several US states — disclosure requirements, data protection obligations, and consumer rights around automated decision-making are all in flux. Second, workforce transition is not being adequately planned for: most hospitality organizations are adopting AI tools without any parallel investment in helping staff understand and adapt to their new role alongside AI systems. Third, quality assurance frameworks for AI-generated content don't exist yet — the industry lacks agreed standards for what "good" looks like, which makes it difficult to hold vendors accountable.
For operators, the practical summary is: use generative AI aggressively in low-risk, high-volume content and communication workflows today — the efficiency gains are real. Build your governance framework for higher-stakes applications (guest data handling, automated communications, pricing decisions) before scaling them. And invest in your team's AI fluency now, while the transition is still gradual enough to manage thoughtfully.