Hotels have adopted AI widely but rarely got the hours back: more than half use or are buying generative AI, yet fewer than one in ten say it cut their manual work by more than 30%. For a travel agency, the lesson is to judge any AI purchase by the manual hours it removes, not by the demo.
Key facts
- More than 270 hotel brands and 58,000 properties are covered — RateGain, The State of Distribution, third edition.
- Fewer than one in ten hotels say generative AI cut their manual work by more than 30% — same report.
- More than 80% of commercial teams spend one to two days a week rebuilding reports by hand; fewer than 30% have invested in dedicated reporting tools — same report.
- 55% of hotels report no change, or only a minor one, to their distribution strategy — same report.
- About 4 in 100 messages in one operations mailbox carry a document, and one in five sits in a folder called "Solving Problems" — the author's operations archive.
The news. The third edition of The State of Distribution, from RateGain with NYU's Tisch Center of Hospitality and HEDNA, covers more than 270 hotel brands and 58,000 properties. More than half of hotels use or are buying generative AI, yet fewer than one in ten say it cut their manual work by more than 30%.
Why was buying the tool never the hard part?
RateGain's founder put the finding in one line: "Buying AI is easy. Getting value from it is not." The report shows where the value is stuck. Hotels trust AI to assist, not to act on its own, and most are spending on fixing systems they already own rather than adding new ones.
That is not caution for its own sake. It is the right instinct. The part of commercial work that carries liability, the rate you commit to and the policy you accept, should stay with a person. The part that carries none is the report someone rebuilds by hand. More than 80% of commercial teams still spend one to two days a week on that, and fewer than 30% have invested in dedicated reporting tools.
What does the same gap look like at an agency?
Travel agencies are about to repeat this, with less data to warn them. The tools on offer are aimed at what is visible: a chat window, a content generator. The hours are somewhere else. In one operations mailbox from my own archive, only about 4 in 100 of the mail carries an actual document. The rest is people chasing people. One message in every five sits in a folder literally called "Solving Problems".
No chat window touches that. The real work is assembly and follow-up: collecting the pieces of an answer from places that were never built to be read together, then chasing the confirmation after the booking. Measure AI there, by the hours it gives back to the desk, or do not measure it at all.
Is AI search already sending bookings?
One more line deserves attention. AI-originated search now produces a measurable share of hotel reservations, yet 55% of hotels report no change, or only a minor one, to their distribution strategy. And online travel agencies still generate close to twice the bookings of hotels' own digital channels.
The intermediary is not disappearing. The open question is which intermediary an AI answer sends the traveller to. For a B2B desk that means what it always meant: be the one whose answer is complete, correct and easy to act on.
Why it matters to the agency: judge any AI purchase by the manual hours it removes, and start with the report or the chase your team repeats every week, not with the tool that gives the best demo.