Corporate buyers want AI to recommend hotels, not book them
Advice, yes. The transaction, no. For hotels, that makes being recommended the whole contest.
GBTA's 2026 study asked corporate travel buyers how comfortable they are with AI making various decisions, and the answers fall along a clean line. Ninety-five percent are comfortable with AI recommending flights and hotels based on negotiated rates. Ninety-two percent will let it generate reports. Then the floor drops. Sixty-two percent will let it execute a change or cancellation. Fifty-seven percent will let it rebook on its own during a disruption.
You can read that as hesitation — buyers aren't ready for autonomous AI yet. Or you can read it as an instruction. AI is welcome to advise, and not to transact, and the line between the two falls right where money moves. Buyers will take the recommendation. They won't hand over the card.
The line consumers already drew
They're not the first to draw it. The consumer market drew the same line in March, and paid real money to mark it.
OpenAI spent most of 2025 building native checkout into ChatGPT — research a hotel, book it in the same window, never touch an OTA. In March 2026 it pulled the feature back into third-party apps. The reason was mundane and damning: people were glad to ask ChatGPT for ideas, but when it came time to enter a card, they left and booked somewhere they trusted. Barely a dozen merchants had gone live before the retreat. Expedia jumped 12 percent on the news; Booking Holdings rose 8. The market had spent months pricing in the death of the intermediary, and the death got postponed.
The relief rally missed the more durable point. The booking didn't move to AI. Discovery did. Travelers now use AI to plan and narrow long before they reach a booking screen, and the contest quietly relocated — from owning the transaction to being the recommendation. Eight months later, from the other side of the room, corporate buyers landed on the same line. Same settlement, different population, no coordination between them.
Why the recommendation is the whole game
If AI recommends and a human books, the recommendation is the part that decides the booking. Obvious enough — until you walk it into a corporate program.
The thing buyers are most comfortable handing AI, by a wider margin than anything else in the study, is recommending flights and hotels on negotiated rates and preferred suppliers. Ninety-five percent. Not a soft lead — the single most accepted use of AI anywhere in the research. So picture the flow the data points to: an agent builds a shortlist around the program's negotiated and preferred content, and the traveler picks from it. The agent narrows. The human confirms. The negotiated rate either makes the shortlist or it doesn't.
Which collapses the corporate and consumer stories into one question. Not "will AI book my hotel." Will AI put my hotel in front of the person booking it. Discovery is the battleground on both sides — the consumer asking ChatGPT where to stay, the corporate traveler handed an AI-built shortlist inside the booking tool. The property that gets recommended gets the stay. The one that doesn't never reaches the part where anyone enters a card.
What it means for a hotel
A hotel can't decide whether a program's AI books on its own. Buyers have already decided that, and for now the answer is mostly no. What a hotel can shape is whether its content is in good enough shape to be recommended.
Which throws the weight back onto something unglamorous: the rate, and what travels with it. An agent recommending on negotiated rates and preferred-supplier status can only weigh what it can read — the rate, the inclusions, the terms, the attributes that make one room a better fit than another. Content that arrives thin, or as a bare number, gives the engine nothing to prefer. The hotel that wins the 95 percent slot is the one whose rate shows up complete enough to be ranked on more than price.
The transaction layer will keep getting argued over — who clicks book, human or machine, and when that flips. It's the louder fight. But the buyers in this study have already pointed at where the decision really gets made, one step ahead of the part everyone is watching. The booking was never the battleground. Being the recommendation is.
by Markus Busch, Editor/Publisher Hospitality.today
Enjoying this analysis? Hospitality.today delivers daily insights on hotel distribution, AI trends, and travel commerce — straight to your inbox. Subscribe for free at Hospitality.today →