The whole industry is building autonomous booking. 7% of travelers want it.

Nine in ten travelers will happily let AI help them find a hotel. Seven percent will let it do the booking — the part the entire industry is racing to build

Look at where the money and the engineering are going. Google is wiring an auction into Gemini's hotel answers. Connectivity platforms are pushing live rates into ChatGPT, Claude and Gemini through new connectors. The chains have shipped their own assistants. A wave of vendors sells faster execution so a hotel can keep pace with machine-speed demand. All of it points at the same destination: the autonomous booking, the moment an agent shops, decides, and buys on the traveler's behalf.

Then look at how many travelers want that. A recent Accenture survey of close to 3,000 consumers puts it at seven percent.

The number worth sitting with is not how cautious people are. It's the gap between what's being built and what's being asked for. Because the appetite for AI in travel is genuinely large — just not where the build-out is aimed. In the same survey, nearly nine in ten travelers are happy to use an AI agent to plan a trip and find options. Help me look is a near-universal yes. Do it all for me is a rounding error. And the middle tells the same story: fewer than a third would even let AI make the final booking decision, and that's with the traveler still handling payment.

So the demand is real and it has a clear shape. It sits at the top of the trip — discovery, comparison, narrowing the field. It thins out fast as you move toward the money. The industry is pouring its effort into the bottom, the commit, the part where the want is thinnest.

There's a reason the want thins, and it isn't only trust. Booking a trip isn't a chore people are desperate to hand off. Accenture found that 44 percent of travelers stay involved on purpose, because they feel a connection to the brand and they enjoy the browsing. A trip is something people look forward to choosing. The autonomous-purchase model was imported from commodity e-commerce — reorder the detergent, refill the dog food, don't make me think about it. A hotel for an anniversary is not detergent. The assumption that travelers want the decision taken off their hands is borrowed from a category travel doesn't belong to.

None of this says autonomous booking won't grow. Seven percent is a floor, not a ceiling, and comfort tends to rise once something works well. The point is about sequence. The race is being run as if the demand for autonomous booking already matches the demand for AI discovery, and it doesn't — one is near-universal, the other is a sliver. The first-mover prize everyone is sprinting toward is a prize most travelers haven't asked to be given yet.

It complicates the agentic story the way a few other quiet facts do. The booking rails still hand off at the OTA. The supply-side connectivity to make any of this seamless is, by the admission of people deep in it, not built at scale. And now the demand for the end state — the agent that buys for you — comes in at seven percent. Three different parts of the same picture, all saying the same thing: the autonomous booking is further off than the noise around it suggests.

Which leaves a more useful question than who wins autonomous booking. It's who serves the nine in ten who want help finding the room and then want to choose it themselves. That demand exists right now, at scale, and it's the part of the journey a hotel can actually compete for — being the option the agent puts forward, and being worth choosing when the traveler, as most of them intend to, makes the call. The booking everyone is racing to automate is the one travelers are least eager to give up.

by Markus Busch, Editor and Publisher of Hospitality.today

Read also: Act faster, rank higher, still book on the OTA

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 →