When a guest asks ChatGPT, Claude or Google's AI search for a «good 4-star hotel by the lake», one small piece of information helps decide whether your property appears in the answer: the star category. This is exactly what the Hotelstars Union addressed at its general assembly in Vilnius on 12 October 2026. Its finding is both encouraging and uncomfortable: large language models already recognise official stars and label them as such. But how they are displayed still depends on the provider, the country and the individual query.
For Swiss hotels, this is not a distant matter. HotellerieSuisse co-founded the Hotelstars Union in 2009, and Swiss stars follow its shared catalogue of criteria.
The problem in everyday hotel business
Classification costs time and money. Every three years, it is reassessed criterion by criterion. Online, the result then sits next to stars that nobody has checked. In many countries, Booking.com shows stars for properties without a recognised classification, sometimes self-assigned, sometimes derived from its own rating. In April 2026, the Hotelstars Union openly criticised this as misleading. Google, in turn, displays its own hotel class, compiled from several sources: partner data, its own research, feedback from hoteliers and machine learning that evaluates price, location, room size and amenities.
To the guest, all these stars look the same. For the hotel, this means: the verified category can appear differently online, depending on which source a system draws it from. With AI assistants, this problem intensifies. A language model does not show a list for the guest to compare. It formulates an answer. If that answer states the wrong category, nobody corrects it.
What the Hotelstars Union now states
The announcement of 12 October gives concrete examples. In its app, Tripadvisor marks properties classified according to Hotelstars Union standards. Expedia names the Hotelstars Union as the source of its stars in several member countries, and in some markets also within AI-assisted search. And via its own connector, official classification data can be queried directly through AI assistants, for example in Anthropic's directory for Claude.
At the same time, the association calls on booking platforms, online travel agencies and providers of language models to use stars only for officially classified hotels and to display them consistently across all channels. The Hotelstars Union comprises 22 member countries and, by its own account, more than 22'000 classified hotels with around 1.2 million rooms.
The core of this reasoning is instructive for every hotel: AI assistants depend on structured, reliable data. An official classification provides exactly that. It is based on fixed criteria, independently verified, and comparable across national borders. Machines favour a feature with these three qualities.
What does this mean for your hotel?
1. Stars belong on your own website in machine-readable form
Many hotels display their stars as an image at the top of the page. A human recognises this, a machine often does not. For this purpose, structured data has a dedicated property: starRating. Schema.org explicitly describes it as an official rating by national associations or standards bodies and recommends naming the issuing organisation via the author property. The definition cites Hotelstars among its examples. Marking up «4 stars, HotellerieSuisse» in this way gives every system a clear answer instead of a graphic.
2. The same category everywhere
Check what circulates about your property: on Booking.com, Expedia, Tripadvisor, in your Google Business Profile, in your booking engine, on the destination's website. Is the category from before your last classification still listed somewhere, or is the «Superior» missing? For its hotel class, Google explicitly also takes hoteliers' feedback into account. A contradictory picture online makes it hard for any system to settle on one answer.
3. Explain the classification in text, not just show it
Language models readily adopt whole sentences. A short paragraph on the website, for example on the «About us» page or in the FAQ, helps more than any symbol: who classified the property, since when, and what the category actually means. Front desk opening hours, breakfast offering, room service: these are facts an AI can use to answer a specific guest question.
4. Use the classification deliberately, not just renew it
Classification is valid for three years at a time. Plan your next renewal so that the website, profiles and platforms are updated on the same day afterwards. Hotels that are not classified at all should reconsider the question: until now, the star was primarily a signal to the guest. Now it is also becoming a data point that machines use to categorise a property.
Common mistakes
- Stars only as a graphic: An image with five stars in the logo tells a machine nothing about the category.
- Mixing review stars and hotel stars: 4.6 out of 5 points on Google is a guest review, not a classification. In structured data, one belongs to aggregateRating, the other to starRating.
- Outdated category in circulation: After an upgrade or downgrade, the old rating often remains on portals and partner sites for months.
- Self-assigned stars: Hotels that display stars without classification risk exactly the confusion the Hotelstars Union is working against. And at a time when systems are getting better at recognising the official source.
- Never asking the systems yourself: Ask ChatGPT, Claude and Google Search what category your hotel has. The answer quickly reveals where an incorrect figure is circulating.
Conclusion
Hotel stars were long a matter for the sign at the entrance and the brochure. With AI search, they are becoming one of the few features that are independently verified and therefore trustworthy for machines. The Hotelstars Union is fighting to ensure that platforms and language models display them correctly. Every hotel can take care of its own part immediately: the category in machine-readable form on the website, the same figure everywhere, and a sentence explaining what it means.
Would you like to know how ChatGPT & Co. describe your hotel today, and whether your classification comes across correctly? We are happy to check this with you: +41 52 566 10 19.
Sources
- Hotelstars Union, 12 October 2026: AI is in love with official hotelstars (press release on the general assembly in Vilnius)
- htr.ch, October 2026: Hotelstars Union nimmt KI-Assistenten in die Pflicht
- Hotelstars Union, 21 April 2026: Call for a Europe wide, consistent and transparent display of hotel star ratings
- HotellerieSuisse: Hotel Kriterienkatalog 2025–2030 (PDF)
- Google Business Profile Help: Hotel class ratings
- Schema.org: starRating
More related articles
- Structured data for hotels: How to stand out on Google with stars, prices and reviews
- Reviews in the age of AI: Why your stars now count twice
- Google Business Profile 2026: What hoteliers need to change so AI can find them
- Why a «beautiful website» is not enough: What AI systems can really read from your hotel website
