78% of chains use AI. Only 41% of independents do. The gap is not what you think.

Two numbers have been circulating since the spring of 2026, usually side by side, usually to make independent hoteliers feel behind.

The first comes from a study by the Institute of Tourism at HES-SO Valais-Wallis, which surveyed 1,485 hotels across six European markets: 78% of hotel chains are using AI, against 41% of independent hotels, with a further 16% of independents planning to adopt it.

The second comes from a global survey by Canary Technologies of over 400 hotel technology decision-makers: 82% expect AI use to increase across their organisation within the next year, and 85% will put at least 5% of their IT budget into AI tools this year.

Read together, the story writes itself: the industry is moving, chains are ahead, independents are lagging. That story is being used to sell a lot of software right now.

It is also incomplete, and the missing piece is the most useful thing in either study.

The number nobody quotes

In the same European survey that produced the 78% figure, only 7% of chains have implemented a company-wide AI strategy.

So the honest version of the headline is this: 78% of chains are doing something with AI. Seven per cent have actually built it into how they operate. The remaining seventy-one points are pilots, departmental experiments, a revenue-management add-on, a marketing team using ChatGPT to write descriptions.

The gap between a chain and an independent hotel is not a gap in capability. It is a gap in the number of experiments running at once.

Why the chains are stuck

The European study names the barrier plainly: integration with legacy systems.

This is not a small thing. A chain with a decade-old PMS, a corporate CRM, brand standards, a procurement process and a legal department cannot simply connect a new tool to its guest data. Every integration is a project. Every project needs a business case, a security review, and a slot in a roadmap. That is why so much chain AI stays in pilot: the pilots are easy, and the rollout is where the system meets the systems.

An independent hotel with 25 rooms has none of that. It also has none of the budget, which is the trade everyone focuses on. But it can decide on a Tuesday and be running by Friday, and no committee has an opinion.

That is a real structural advantage, and it is the first time in twenty years of hotel technology that scale has been the disadvantage.

What independents are actually adopting

The same research found that independent operators are not attempting the ambitious projects. They are adopting practical, guest-facing tools: content generation, chatbots and messaging, review analytics.

There is a pattern in that list. Every item is narrow, visible, and does not require touching the PMS. They are the tools where you can tell within two weeks whether it worked, because either the messages are being answered or they are not.

This is the opposite of how chain AI programmes usually start — with a data project, a warehouse, a model, and a benefit that will be measurable next year.

The independents are, accidentally, doing it in the right order.

The three objections, and where each one now stands

"It costs too much." This was true and is no longer. Guest messaging tools that cost thousands a month in 2023 now start in the low hundreds, and the underlying model cost has fallen by more than an order of magnitude. The real cost of an AI assistant for a small property is not the licence — it is the two days someone spends writing down what the property actually knows. That cost has not moved, and never will.

"We don't have the skills." Also true, and also less true than it was. The skill required is no longer technical. It is the ability to describe your own property precisely: where guests park, what the ZTL rules are on your street, what time the bins go out. That is the job of whoever runs the property, and nobody else can do it.

"We're worried about the data." This one deserves respect rather than dismissal, because it is the objection that will still be valid in three years. A guest's passport, arrival time and room number are personal data, and under the GDPR the hotel is the controller — responsibility does not transfer to the vendor. Under Article 50 of the EU AI Act, applicable since 2 August 2026, a system that talks to guests must be designed so they know it is not a person.

But the conclusion people draw from this is backwards. The risk is not in adopting AI. It is in adopting it without asking the vendor where the data goes, whether a processing agreement exists, and how long conversations are kept. Those are three questions, not a research project. A vendor who cannot answer them in an email is the actual risk.

What the gap means commercially

There is a version of this argument that ends with "so independents should hurry up". That is not the conclusion.

The useful conclusion is narrower. Chains are running many shallow AI experiments, most of which will not survive contact with their own IT estate. Independents are running few, but the ones they run are guest-facing and go live. Over the next two years the difference in outcomes will be much smaller than the difference in adoption rates suggests — and in the specific area of guest communication, where the small property has fewer systems in the way, the independent may well end up ahead.

That is worth saying to any hotelier who has been shown the 78% and told they are behind. They are behind on pilots. Pilots are not the thing.

Where to start, if you are in the 41% who haven't

Pick the narrowest problem you have that is (a) repetitive, (b) currently answered by a person, and (c) has a fixed correct answer. For most independent properties that is inbound guest messaging: the same questions about parking, arrival, wifi and luggage, arriving at all hours in several languages.

Write down the answers properly — not as a policy document, as you would say them out loud. That document is the actual asset. The tool on top of it is replaceable.

Then measure one thing for a month: how many messages a person still had to answer. If the number does not fall, the information was too thin, not the technology.

---

Published July 2026. Figures: Institute of Tourism, HES-SO Valais-Wallis (1,485 hotels, six European markets) and Canary Technologies, "Navigating AI: Hospitality Shifts From Exploration to Execution" (400+ hotel technology decision-makers, global). We review this article every six months.