Strategy
From SEO to GEO: what AI search changes for an independent property
Roughly half the properties an AI assistant recommends to a traveller do not appear anywhere in Google's organic results for the same question.
That figure comes from 695 unbranded hotel searches run through ChatGPT and Google's AI Mode, published on 24 June this year. The queries were the kind a traveller actually asks: family hotels in Mallorca, where to stay in Rhodes. Between 48% and 56% of the recommended properties were absent from the organic results.
Ten years of search advice assumed that if you rank, you get found. The lists now being handed to travellers are drawn from somewhere else.
What is GEO, and how is it different from SEO?
Generative engine optimisation is the practice of getting a property named inside an answer that an AI assistant writes, rather than ranked in a list of links. SEO competes for a position on a page of results. GEO competes for a place in a shortlist of three or four names, where there is no second page.
The disciplines overlap more than the acronym suggests. Both reward clear pages, real authority and content a machine can read. What changes is the shape of the result. A page of ten blue links gave every property in the top ten a chance at the click. An assistant that names four lodges has quietly disqualified everyone else before the traveller has read a word.
| SEO | GEO | |
|---|---|---|
| What you win | A position in a list of links | A named place inside a written answer |
| How many can win | Ten on the first page, more behind it | Three or four, and there is no second page |
| What decides it | Your rank | What sources across the web say about you |
| Where the reader goes next | Your site, if they click | Often nowhere. The answer was the destination |
| What you control | Your own pages | Your own pages, plus your accuracy everywhere else |
Does ranking on Google still get you recommended?
Yes, and less reliably than it used to. The same June study measured how often a property at each Google position was named by an AI assistant.
| Position in Google's organic results | Named by an AI assistant |
|---|---|
| Top 3 | 66% |
| Top 10 | 61% |
| 11 to 20 | 45% |
| Below 20 | 27% |
So rank still helps. It is no longer the gate. A property sitting at position 15 gets named nearly half the time, and a property Google does not rank at all can still turn up in the answer, because the assistant assembled its list from sources that were never your website.
Where do the assistants actually get their answers?
Mostly from other people's sites. Across those 695 searches, a hotel's own domain was cited as a source less than 10% of the time. The citations went to online travel agencies, review platforms and editorial guides.
A separate study by Cloudbeds, published in June 2025, put a number on the same pattern from a different angle. Across hundreds of automated queries in six destinations, run through ChatGPT, Perplexity and Gemini, more than half of all cited sources, 55.3%, were OTAs. Booking.com, Expedia and Tripadvisor led them.
There is an uncomfortable reading of this, and it is the correct one. A property can spend years and a large budget fighting the OTAs for the direct booking, and the assistant now deciding which properties a traveller considers is reading the OTA listing to make up its mind.
Cloudbeds returned to the subject in its 2026 State of Independent Hotels Report, drawn from 90 million bookings across roughly 180 countries, and put OTA share of independent hotel bookings at 63.4%. The channel an independent property spends most of its energy trying to escape is also the channel teaching the assistant what that property is.
The Cloudbeds work also mapped where recommended properties showed up elsewhere: 98% appeared on YouTube, 97% in travel blogs, 95% on Reddit. None of those are places a marketing team controls. All of them are places a marketing team can be present in, accurately, if someone is paying attention.
Why does this hit independent properties hardest?
Because the assistants favour the groups. In the Cloudbeds study, 72.4% of AI-recommended hotels were branded properties or part of large groups. Independents trailed by 4.43 percentage points on visibility.
Behind that sits a wider gap in who is even working on it. The 2026 report found nearly 80% of hotel chains using AI in some form, against 41% of independents.
That visibility gap is not enormous, and it points the wrong way for most of the properties we work with. A single owner-run camp has no brand estate, no group PR budget and no corporate domain accumulating mentions. What it usually has instead is better reviews, a more specific story and an actual reason to be chosen, none of which is doing any work if the property never makes the shortlist.
Worth sitting with: recommended properties in that study averaged a sentiment score of 75 out of 100 across review platforms, with high review volume. Reputation is the one input on this list where a small property can genuinely beat a group, and it is largely earned rather than bought.
Is there a single answer to optimise for?
No, and this is the part most coverage skips. In the June study, ChatGPT named 3,151 distinct hotels. Google's AI Mode named 1,286. Only 691 properties appeared on both, an overlap of about 18%.
Two assistants asked the same question about the same destinations produced substantially different worlds. There is no equivalent of "the number one result" to aim at. What there is instead is a set of conditions that make a property legible and credible to any of them, which is a slower and less satisfying thing to sell than a ranking.
How many travellers are really doing this?
Fewer than the loudest numbers suggest, more than most owners assume, and the honest answer is that nobody knows precisely.
Phocuswright's 2026 research found 39% of US travellers actively using AI to plan trips, up from 28% a year earlier, and 56% using it for planning, booking or in-destination help on at least one trip in the past twelve months. An IMG survey of just over 1,000 people in March 2026 found 33% likely to use AI tools for 2026 travel planning. TakeUp surveyed 300 US leisure travellers in April 2026 and found 38% had used AI for trip planning. Accenture, asking 18,000 consumers across 14 countries a broader question about AI tools in travel, reported 80%.
Those studies are not measuring the same thing, and any article quoting only the 80% is selling something. The consistent finding across all of them is direction rather than level: the share is rising every time somebody measures it.
The more useful Phocuswright figure is about where research now starts. General search engines fell from 51% to 36% as travellers' most-used research resource between mid-2024 and late 2025, while use of generative AI platforms nearly tripled over the same period. And 44% said they would book directly inside an AI platform if they could.
What does this do to the traffic you already have?
It thins it. When a Google AI Overview appears, click-through rate for the top-ranking result falls by about 58%, according to Ahrefs analysis of December 2025 data. Seer Interactive tracked 3,119 informational queries across 42 organisations between June 2024 and September 2025, covering 25.1 million organic impressions, and measured click-through on those queries falling from 1.76% to 0.61%. Pew Research Center put browsing trackers on more than 900 US adults in March 2025 and logged 68,879 Google searches: people clicked a traditional result on 8% of visits where an AI summary appeared, against 15% of visits where none did.
A property can hold its rankings and watch its organic traffic fall, and there is nothing wrong with the website. This is the year a lot of owners will be shown a graph going down and told the SEO stopped working.
The same Seer analysis found the other half of it. Brands cited inside an AI Overview earned 35% more organic clicks. The traffic is not only thinning, it is concentrating on whoever gets named.
What actually helps
None of this is exotic. Most of it is unglamorous work a small team can do without an agency.
- Get the facts out of PDFs and booking widgets. Rates, what is included, transfer times from the airstrip, the malaria position, whether the paths take a wheelchair, minimum stays, child policies. An assistant cannot quote a rate sheet it cannot open, and much of the information that decides a booking sits inside a downloadable PDF or renders inside a booking engine.
- Publish the questions your reservations team already answers by email. Every question answered privately six times a week is a page that would earn the enquiry instead. This is the cheapest work on the list and it does two jobs, because those pages are exactly what an assistant quotes.
- Check who you are blocking. GPTBot, ChatGPT-User, PerplexityBot, ClaudeBot and Google-Extended are the crawlers behind the assistants doing the recommending. Plenty of properties have them disallowed in robots.txt without anyone having decided to block them. A blocked crawler cannot cite you.
- Be consistent about who you are. The same property name, the same address, the same description, everywhere it appears. Assistants assemble an understanding of a property from fragments across the web, and a lodge listed under three slightly different names is three weaker entities rather than one strong one.
- Look after the third-party surfaces, since that is where the citations come from. Review platforms, OTA listing content, the travel blogs and YouTube channels that cover your region. You are not going to control them. You can make sure they are accurate.
- Cite things when you write. In the Princeton GEO study presented at KDD 2024, adding citations to sources raised a page's visibility in generative answers by up to 40%, and adding statistics by 37%. Keyword stuffing reduced it by 10%. The gains were largest for sites with the least existing authority.
The part the data does not cover
Every study quoted here looked at hotels. The Tharro searches ran in the Algarve, Mallorca and Rhodes. A four-night beach hotel booked six weeks out is not a $9,000-per-person tented camp booked eleven months out after four conversations, and we do not know whether these numbers hold at the top of the market. Somebody spending that much keeps researching long after the shortlist forms, which argues that the assistant matters less here than it would for a weekend booking. The counter-argument is that the shortlist is where the decision gets framed, and a property that never enters it is never in the running to be researched at all.
Nobody has run this study on African safari lodges. Until somebody does, anyone telling you exactly how AI search behaves in this segment is guessing, ourselves included.
What we would say with more confidence is that the assistant decides who gets considered, and it does not decide who gets chosen. Being named in a shortlist of four is the beginning of the work. The property still has to be the one that is wanted, and no amount of machine-readable structure has ever made anybody want to go somewhere.
Common questions
Is SEO dead?
No. Ranking still correlates strongly with being recommended, and the technical work that helps a search engine also helps an assistant. What has changed is that ranking is no longer sufficient on its own.
Should we block AI crawlers to protect our content?
Only if you have decided you do not want to be recommended. Blocking GPTBot or PerplexityBot removes the property from those platforms' ability to cite it.
Do we need schema markup?
Google states that no special markup is required for its AI features. Other assistants do reward structured, extractable content. Schema is worth having and it is not the thing standing between you and a recommendation.
How would we know if this is working?
Ask the assistants directly, on a fixed set of queries, once a month, and write down what comes back. It costs nothing and it is more informative than any dashboard currently available.
Waeve is a performance marketing studio for luxury travel. We run the ads, build the sites and shape the sales process for a small number of properties each year.
Field notes go out as we write them. waevedigital.com
Sources
- Tharro, AI hotel visibility research, 24 June 2026. 695 unbranded hotel searches across the Algarve, Mallorca and Rhodes, run through ChatGPT and Google AI Mode. hotelnewsresource.com
- Cloudbeds, The Signals Behind Hotel AI Recommendations, 26 June 2025. Hundreds of automated queries across six destinations, 145 consistently top-ranked properties, via ChatGPT, Perplexity and Gemini. cloudbeds.com
- Cloudbeds, 2026 State of Independent Hotels Report, 25 March 2026. 90 million bookings across roughly 180 countries, tens of thousands of properties, covering 2025 performance. cloudbeds.com
- Phocuswright, 2026 traveller research on AI adoption and research channels. phocuswire.com
- Ahrefs, AI Overview click-through analysis, December 2025 data. medianama.com
- Seer Interactive, AIO impact on Google CTR, Tracy McDonald, 4 November 2025. 3,119 informational queries across 42 organisations, 25.1 million organic impressions, June 2024 to September 2025. seerinteractive.com
- Pew Research Center, Google users are less likely to click on links when an AI summary appears in the results, 22 July 2025. Browsing-tracker panel of 900+ US adults, 68,879 Google searches, March 2025. pewresearch.org
- IMG Travel Outlook Survey, March 2026. prnewswire.com
- TakeUp, AI travel planning research, April 2026. takeup.ai
- Accenture, Consumer Pulse survey, reported 9 June 2025. 18,000 consumers across 14 countries. This is the oldest of the four adoption figures quoted and the largest. globetrender.com
- Aggarwal et al., GEO: Generative Engine Optimization, KDD 2024 (Princeton).
- Google, AI features and your website. developers.google.com
