Last spring, I stayed at a hotel near a conference. I can't recall the name without checking my email, which tells you something. It was fine. That is the word that kept coming to mind. The room was what I expected, the kiosk worked, and nothing went wrong. I did not think about it again until I was writing this piece, and even then, I had to look up the booking confirmation to remember the property at all. The word 'fine' is dangerous for hospitality. It means the guest left feeling nothing and will book somewhere else next time without quite knowing why.
That is not a neutral outcome. That is a failure; the property was never registered because nothing went wrong. Nothing went right either.
I keep coming back to this when I think about what the hotel industry is measuring right now and what it is missing. Check-in time. Chatbot containment rate. Ticket resolution speed. These numbers are real. The efficiency gains behind them are real. But they measure the transaction, not the experience. And guests, it turns out, remember the experience.
The numbers, on the other hand, are worth sitting on. A University of South Florida study found what researchers described as a striking gap between how hotel managers and guests see AI adoption; managers viewed it as a genuine relief from workload and staffing pressure, while guests were far less convinced. Hotels.com ran a survey last year, of 450-plus properties, global sample, results published September 2025, and the number that stuck with me was this: 70 percent of hotels said their guests still prefer talking to a person, especially at check-in and when things get complicated. These are not technophobic guests. Many use AI tools regularly in their work. Their hesitation is not about capability. It is about what a hotel stay is supposed to feel like, and they quickly know when it does not.
The USF study gets more specific: 81 percent of respondents identified emotional authenticity as a core problem with AI-based services, and guests consistently wanted human staff for requests with any emotional weight. An anniversary dinner reservation. A noisy room at midnight. A situation with no obvious solution. Not because an AI concierge gives the wrong answer, often it gives a technically correct one, but because, in those moments, what the guest needs is not information. It is the sense that someone in the building decided her situation was worth solving personally. That is a different thing entirely.
The luxury market is where this gets genuinely uncomfortable to think about, and the industry hasn't sat with it long enough. The properties investing most aggressively in AI tend to be the same ones whose guests have flown business class to get there, have stayed in that brand 40 times, and expect, not unreasonably, that someone will remember their name without being prompted by a screen. There is something almost paradoxical about that, except it is not really a paradox; it is just two trends moving in opposite directions, and nobody is yet willing to say so out loud.
A guest paying $1,000 a night has already assumed the room will be ready and the check-in will be smooth. She is not grateful when those things work; she notices when they do not. What she will remember, and tell her colleagues about, is harder to track. The front desk agent, who noticed she looked worn out, quietly moved her to a higher floor. The concierge recalled, from a stay a year and a half ago, that she does not like sitting near open kitchens. These things are not scalable in any conventional sense. They are, however, the point.
None of this is an argument against AI. That framing is a distraction. The real questions are: Which problems AI is suited for, which it handles poorly, and whether the hotel industry has been too imprecise in its distinction. There are things AI does well that honestly nobody misses doing by hand, such as the 3am inquiry about checkout time, the pricing audit that used to eat a Friday afternoon, and keeping track of which guests requested a foam pillow on their last visit, so someone does not have to dig through notes before they arrive. These are not glamorous problems. They are just problems, and if a system handles them reliably, the staff who are used to handling them can be somewhere else, somewhere that actually requires a person to be present and pay attention.
Where things go wrong, it is harder to articulate, but I think most experienced hoteliers recognize it immediately. It is when AI starts handling the parts of a guest's experience that carry emotional weight, complaints, special requests, moments of genuine friction, and the response, however technically accurate, feels generated rather than chosen. Guests can tell. Not always consciously or immediately, but they can tell. The gap between a message written for this specific person and one written for someone who statistically resembles her is smaller than it used to be, and guests are more attuned to it than most hotel operators realize.
The framing I find most useful is not AI versus people, but AI freeing people to do the things only they can do. A front desk agent who has not been processing check-ins manually for six hours has time to notice the guest who seems off. A concierge whose screen already shows a guest's history from three previous stays can skip the generic questions and get to what matters. Technology, in this model, is not the experience; it is what clears the path for the experience to happen.
That distinction keeps getting lost, and I am not sure it is getting lost by accident. Efficiency is easier to budget for than presence, and what gets measured tends to be what gets funded. Which brings me to something I ask my students every semester, partly because it cuts through a lot of noise. After a stay, any stay, budget or luxury, business or leisure, can you remember a specific person? Not the hotel. Not the app that sent you a pre-arrival message. A person. Someone who said or did something that stuck.
A name, maybe, or just a face, a moment where someone solved something, or noticed something, or made you feel briefly that your comfort was worth someone's full attention. That is what guests carry with them. It is what brings them back. And right now, it is the thing most at risk of being optimized away, quietly, efficiently, and without anyone realizing what was lost.
Jun Kwon, Teaching Instructor, School of Hospitality Leadership, East Carolina University
