Travellers are far readier to let an AI agent compare than to let it pay: in Accenture's survey of more than 5,000 travellers across 12 markets, 40% would fully delegate comparing travel options, but only 10% would fully delegate payment. The commitment, and everything that follows it, is the part an agency can own.
Key facts
- 40% of travellers would fully delegate the comparison of travel options to an AI agent; only 10% would fully delegate payment execution — Accenture.
- More than 5,000 travellers surveyed across 12 markets, the UAE and Saudi Arabia among them, plus 600 senior executives — Accenture.
- 30% of the executives say their organisations are fully ready to run end-to-end journeys across partners — Accenture.
- More than 410 individual hotels chased directly in one desk's correspondence — the author's operations archive.
The news. Accenture's report "Win the AI agent, win the traveler" rests on two surveys run in June: more than 5,000 travellers across 12 markets, the United Arab Emirates and Saudi Arabia among them, and 600 senior executives at airlines, hotels, and online travel agencies and platforms. The traveller answers split sharply. 40% would fully delegate the comparison of travel options to an AI agent. Only 10% would fully delegate payment execution. On the industry side, just 30% of the executives said their organisations are fully ready to run end-to-end journeys across partners.
Why will travellers let AI compare but not pay?
The gap between those two numbers is the gap between looking and owning the outcome. A comparison can be wrong and nobody loses anything until someone pays. The moment money moves, the traveller wants someone who answers for it.
The report makes the point directly: "For travel, with its pricing volatility, cancellation rules, multi-currency payments and post-booking servicing, the execution bar is higher still."
What does a desk actually do after the payment?
Everything after the payment is where the work sits: the rate that changed, the confirmation that did not come, the misspelt name, the room type the hotel says it never sold. None of it is solved by a comparison. It is solved by someone who can reach the hotel. In the correspondence of one desk I worked on, more than 410 individual hotels were chased directly, one property at a time.
A comparison ends when someone presses pay. Execution starts there. That is why being the cheapest line on a list is not enough for an agency. It has to be the one people call when something goes wrong.
Why does it matter for the agency?
Travellers are telling the industry where they will let a machine in and where they will not. Let it compare; that part of the job is leaving. Own the part they are keeping back: the commitment and everything that follows it. An agency that can say "we paid, the hotel confirmed, and if it changes we fix it" is selling exactly what the survey says travellers are least willing to hand to a machine.
Source: Accenture — Win the AI agent, win the traveler (report, PDF). Summary coverage: Hotel News Resource.