When a traveller asks an AI “where should I stay in the Alps for a family ski week”, one hotel gets named in the answer. The others are not in the room. There’s no page two to scroll to and no list to climb back up: you’re either in the answer or you don’t exist for that traveller.
So the question that decides whether that traveller ever reaches you is no longer “does my page rank”. It’s “when the AI answers a question my hotel could win, is my name in the answer, or my competitor’s”.
I looked at five hotels to find out what actually gets a property named. The pattern was consistent, and it’s not the one most hotels are being sold.
The alpine ski group, June to July 2026
RANKS HIGHER IN NORMAL SEARCH
- Rooms and room types
- Booking and availability
- Homepage
- The “our hotels” hub
NAMED IN THE AI ANSWER
- Ski-school guide
- Which village for beginners
- First week on snow with kids
- Family-programme pages
What I looked at
One of the five is an alpine hospitality group with enough traffic to measure properly. Across June and July 2026 it generated about 8 million normal search impressions and roughly 1 million impressions inside Google’s AI answers (AI Overviews and AI Mode), spread across 891 pages that showed up in both. This is measurable at all because Google began rolling out a separate Search Console report for its generative AI features in June 2026, so for the first time you can see which of your own pages Google’s AI is actually surfacing, instead of guessing from traffic.
The other four are smaller, the kind of single-property or independent sites most hotels run. I ran the same comparison on each, lining up the pages that rank in normal search against the pages that get named in AI answers. I’m not publishing their numbers, partly because they’d be identifiable and partly because four small sites shouldn’t be turned into a law of how all hotels behave. They’re here as confirmation of a pattern, not as a second dataset.
The pattern: you get named for what you’re known for
Across all five, the pages that got named in AI answers were not the best-optimised pages, and often not even the highest-ranking ones. They were the pages tied to whatever each hotel was already known for. Here is what that looked like, kept at the level of page types rather than individual URLs.
| Hotel | Known for | Named in AI answers | Ranked higher, missed | Verdict |
|---|---|---|---|---|
| Alpine ski group | Family ski weeks, ski school | Ski-school and village guides | Rooms, booking, homepage | Match |
| Family beach resort | Kids club, all-inclusive | Kids-club and family-suite pages | Gallery, offers, brochure | Match |
| City boutique | The landmark, the spa | Walking-time and spa pages | Rooms, design lookbook | Gap known for, versus the brand story |
| Airport hotel | Access, short stays | Shuttle, layover, invoice pages | Meeting rooms, corporate rates | Break the commercial pages lost |
| Wellness hotel | Quiet, spa, hiking | Spa-ritual and hiking pages | Ski-pass and winter packages | Match rankings lagged |
The alpine group (Hotel E). Known across the web as a family-ski operator, for its ski school and week-long family programmes, not for luxury design or nightlife. The pages that got named were the ski-school guides, the “first week on snow with kids” explainers, the family-programme pages, the “which village is gentler for beginners” content. Several of those ranked outside the top 10 in normal search and still pulled heavy AI visibility. The pages that ranked higher and barely showed up in AI answers were the room-type pages, the booking and availability pages, the brand homepage, and the generic “our hotels” hub. The match is clean. The web already treats this group as the family-ski operator, and the AI visibility sat exactly on that topic.
The family beach resort (Hotel A). All-inclusive, big kids club, known for family beach holidays and on-site childcare. Named in AI answers: kids-club hours, “what’s included for children”, family-suite explainers, “is it good for toddlers or teens”. Ranked higher but weak in AI: the photo-gallery rooms pages, the offers and booking pages, the “the resort” brochure page. The match is just as clean.
The city boutique (Hotel B). Two streets from a major landmark, with a small spa. The owner wants to be known as a design hotel. The web knows it for the location and the spa. Named in AI answers: “walking time to the landmark”, the neighbourhood guide, the spa-on-a-short-break pages. Ranked higher but weak in AI: the rooms pages, the “boutique design” lookbook, booking. This is the first useful gap. The hotel talks about design, but the AI pages and the wider web both talk about location and spa. That’s not a failure of the pages, it’s a signal that the story on the website is not yet the story the market repeats.
The airport hotel (Hotel C). A business and airport hotel at a northern European hub, known for short stays, airport access, and reliable desks and invoicing. This one broke the tidy version of the pattern, and it’s the most instructive of the five. Named in AI answers: airport-transfer time, “can I sleep six hours between flights”, late-checkout and early-check-in pages, invoice and VAT explainers. Ranked higher but weak in AI: the meeting-room capacity pages, the corporate rates, the homepage. The sales team wants conferences and events, and there’s real organic demand for meeting rooms, so those pages rank. But the traveller asking an AI about this hotel is asking “can I make the morning flight and still get some sleep”, not “how big is your ballroom”. The hotel is known for access, and the AI answers named it for access, regardless of what the commercial team wishes it were known for.
The wellness hotel (Hotel D). Adults-oriented mountain wellness hotel in the same broad alpine region as the ski group, much smaller, trying hard to escape the family-ski association of its valley. Named in AI answers: the spa-ritual and quiet-stay pages, the “hiking from the door” guides. Ranked higher but weak in AI: the ski-pass and winter-package pages, which the site still ranks for simply because it sits in a ski valley and inherits that demand. In winter, normal search keeps pushing ski intent onto those pages. The AI visibility stayed on spa and quiet anyway. The thing the hotel wants to be known for held in the AI answers, even while its rankings pulled toward a product it’s trying to leave behind.
So: a clean match in three of the five, a known-for versus brand-story gap in Hotel B, and a genuine break in Hotel C where the higher-ranking commercial pages were simply not the pages the AI named. In every case, including the two that broke, the AI visibility tracked what the hotel is actually known for, not what ranked highest and not what the owner wanted to sell.
Why the ski-school page beats the rooms page
Here’s the mechanism, and I want to be honest that it’s the best available explanation rather than something Google’s report proves.
When a traveller asks an AI a question, the AI doesn’t go looking for one answer to that one question. It quietly breaks the question into a handful of smaller ones and answers all of them at once. “Best family ski hotel in the valley” becomes, underneath, a set of questions: is there a ski school for first-timers, can you ski straight in or is it a bus ride, are there rooms that sleep four, what are the kids-club hours, is January or March better, which village is gentler for beginners.
Your rooms page can rank first, because by the time someone searches your rooms, they’ve already typed your hotel’s name. But the AI is usually trying to finish the trip, not match your name. So it reaches for the page that answers “ski school for first-timers” and “which village is gentler”, and that page is your ski-school guide, not your rooms page. The page that wins is the one that answers the traveller’s real question, even when it ranks lower.
That’s why, across the five, the questions that decide whether a hotel gets named look like this:
- The ski group lives or dies on “ski school for first-timers”, “ski-in or bus”, “rooms that sleep four”, “which village is gentler”.
- The beach resort lives on “toddler or teen”, “what’s actually included”, “distance to the beach”, “is there evening childcare so a parent can disappear for two hours”.
- The city boutique lives on “walking minutes to the landmark”, “quiet rooms”, “is the spa real or just a treatment room”, “which side of the square to stay on”.
- The airport hotel lives on “shuttle time”, “earliest check-in”, “shower and bed for a layover”, “a desk if the layover turns into a work day”.
- The wellness hotel lives on “adults-only or not”, “hiking loops from the door”, “how long the treatments run”, “is the village a ski circus in February”.
The honest limit: Google doesn’t tell me which of these questions triggered each AI appearance, so I can’t prove the breakdown question by question. What I can see is the shape of the result, which is that the trip-answering pages get named while the higher-ranking commercial pages don’t. This is the mechanism that explains that shape. Treat it as the best explanation, not as something the report handed me.
Why the tactics you’re being sold don’t fix this
This is where most “AI optimisation” advice falls apart. Turning every heading into a question, bolting FAQ boxes onto every page, writing “citation-ready” paragraphs, none of it changed which pages got named. Those things tidy a page. They don’t change what your hotel is known for, and being known for something is what the AI is rewarding. You can spend the entire budget polishing pages and stay out of the answer.
The thing that put the alpine group in the answer wasn’t clever formatting. It was that the wider web already treats that group as the family-ski operator, in its own content, in what it ranks for, and in what other people write about it. That association exists beyond any single page, and no page trick manufactures it. A page can be improved. What a hotel is known for has to be earned, repeatedly, across everything the web says about it.
When being absent from the AI answer is fine
Reading low AI visibility as failure will send you fixing things that aren’t broken. When someone searches your hotel by name, or lands on your booking page ready to book, there’s no open trip question for the AI to finish, so those pages will sit low in AI visibility, and that’s correct. A booking page does its job at the point of sale, not at the point of “where should I go”. The pages worth worrying about are the ones tied to what you want to be known for, that answer a real traveller question, and still aren’t showing up. That gap is a problem. A booking page with low AI visibility is not.
If you own the hotel, or you’re buying one
If you own the hotel, the most useful thing this data gives you is not another score to chase. It’s a read on whether the market already believes the story you think you’re selling.
Hotel C’s meeting-room pages rank, and still lose the AI answer to its airport-access pages, because travellers ask an airport hotel about sleep and shuttles, not conferences. Hotel B can invest in a design-hotel narrative and still get named for the landmark next door, because that’s what the web repeats about it. That gap is the real question: are you selling the story on your website, or the story the web already tells about you? When those two don’t match, the AI sides with the web, every time.
For anyone buying or repositioning a property, the same read becomes an underwriting question. A family-ski association the web already repeats is a more durable demand signal than a number-one rooms page, because it keeps working when the traveller doesn’t yet know the hotel’s name. It pulls demand at the “where should I go” stage, not just the “book the place I already chose” stage. A repositioning, whether toward boutique design, conferences, or adults-only, will not show up in AI answers until the wider web starts repeating the new story, and ranking the new pages is not enough on its own. And two hotels in the same valley are not really comparable if one is who the AI names for “family first week” and the other is who it names for “quiet spa”, even when their rooms and rates look identical on a spreadsheet. You’re buying a brand-topic association, or the absence of one, and it doesn’t appear on the P&L.
How I checked
I took the pages that appear in Google’s normal search report and the pages that appear in its generative AI report, kept the ones that show up in both, and compared how much AI visibility each page pulls against how much normal search traffic it pulls. Pages that punch above their weight in AI answers, relative to their organic size, are the ones worth studying, and in the measured property the highest-ranking commercial pages accounted for only a small share of AI visibility. Google caps each of these reports at 1,000 pages and does not reveal the question behind each AI appearance, so this is a strong signal rather than a full census, and it’s specific to Google’s AI answers. ChatGPT, Perplexity, Gemini and Claude work differently, and none of this should be read across to them.
See also: Is SEO dead for hotels? | How Google AI Overview picks hotels | Why your hotel is not in Google AI Overviews | The hotel AI visibility audit | The agent will not ask for the best hotel
Questions this study answers
Which of my hotel’s pages get named in AI answers?
The pages that answer a traveller’s real trip question, not your rooms or booking pages. Across five hotels, a ski group was named through its ski-school and which-village guides, a beach resort through its kids-club pages, and an airport hotel through its shuttle and layover pages, because the AI is trying to finish the trip rather than match your name.
Does adding FAQ schema get my hotel into AI answers?
On its own, no. Question headings, FAQ boxes and citation-ready paragraphs tidy a page, but they do not change what your hotel is known for, and that association is what Google’s AI answers rewarded across all five hotels. You can spend the whole budget on formatting and stay out of the answer.
Why is my booking page not in AI answers?
Because there is no open trip question left for the AI to finish. Someone who reaches a booking page has already chosen the hotel, so that page does its job at the point of sale and sits low in AI visibility, which is correct. The pages to worry about are the ones tied to what you want to be known for that still do not appear.