Original travel research, by the numbers
Reports built from named, dated datasets: a running monitor of where AI answers cite this site, readings of the sources ChatGPT cites, proprietary AI citation logs, 62 million search impressions, seven-country trend exports and verified third-party research. Every figure carries its source and its date. If a number appears here, you can trace it.
How this series works: three kinds of sources, always identified in the report. Proprietary first-party data (our citation monitor, built from the AI Performance report in Bing Webmaster Tools, our readings of the sources ChatGPT cites, and our AI citation and grounding-query logs). Large public datasets (Google Trends co-search exports, Bing keyword impressions), stamped with their capture date. And named third-party research (Pew Research Center, Fractl × Search Engine Land, Google’s AI Mode disclosures, Alphabet SEC filings, and the published reports of UN Tourism, the OECD, Eurostat, Deloitte, Mastercard, American Express, Tripadvisor and McKinsey). Where a dataset has a blind spot, the report says so in the copy.
The citation monitor
Where AI answers cite this siteHow often AI answers cite a travel publisher
Every page on this site that is cited in AI answers on Microsoft Copilot and its partners is logged in Bing’s AI Performance report. In the ninety days from 2 July to 29 September 2026 the dashboard counted 418.1K citations. The daily figure reached 15.2K on 29 September, across 291 pages, against an average of about 4,600 a day over the ninety days.
The page-level export, which runs to 28 September, lists 478 pages cited, grouped here into 28 content clusters. Six clusters carry most of the citations: destination rankings (87,575), flight compensation and EU261 (48,769), travel credit cards (47,460), airlines and cabins (43,849), safari and expeditions (32,137) and activity booking platforms (30,310).
Volume is one measure and share is the other. On tour operator questions, pages on this site made up 37% of every source cited; on flight compensation 35%, and on airlines and premium cabins 32%.
The export runs a day behind the dashboard and reads a little lower, so the cluster counts add up to less than the dashboard total. Bing describes its AI Performance data as a sample of overall activity. The next reading comes with the next export. For the questions the assistants ask when they cite the site, see What travellers ask AI.
The ChatGPT readings
Which sources ChatGPT cites on safari questionsOn 50 luxury safari questions, ChatGPT cites this site more than any other
Bing’s report covers Copilot and its partners, not ChatGPT, so ChatGPT is read separately. A fixed set of 50 luxury safari questions is put to it once a day, and every source in each answer is recorded.
Across the 300 answers from 23 to 28 September 2026, pages on this site made up 10.5% of all sources cited, ahead of africansafarimag.com (4.8%) and safaribookings.com (4.6%). This site was the most-cited source on 15 of the 50 questions and among the six most-cited on 32.
An earlier one-off reading, on 6 September 2026, put 40 safari-buying questions to ChatGPT, each in a new, logged-out chat so that no history could shape the answer. This site was cited in 8 of the 40 answers, 23 times in all.
The reading describes these 50 questions, not everything ChatGPT is asked. It counts sources cited, which is different from the companies an answer names; that is what the AI Visibility Report below measures.
The AI visibility reports
Who the assistants name, measured quarterlyWho the AI assistants name for a luxury safari
A traveller planning a safari now asks ChatGPT, Copilot, Gemini or Claude which operators to trust, and gets a shortlist of four or five names with sources. This report measures who is on that shortlist.
Each quarter we put 24 buyer questions to all four assistants, three times each, and record which of 30 leading luxury safari operators are named in each of the 288 answers. The chart ranks the 30 by the percentage of answers in which they were named. No operator can pay to be on it, and no operator can pay to be left off it.
The first chart is published on this page in October 2026, with the dates of the reading printed beneath it. The same questions go back to the assistants every quarter.
Share of voice in AI answers measures presence, not quality. It is separate from the editorial Luxury Safari Operator Index, which is scored on published evidence. Method and definitions are in the questions at the foot of this page.
The machine layer
How AI changed travel discovery
What travellers ask AI: 17,600 questions analysed
Our own citation logs, opened: 576 grounding queries, flight compensation and safaris taking half the pie, and barely 1% phrased like a human.
July 2026
AI vs the click: travel search by the verified numbers
68% of US searches now end without a click; the Fractl million-keyword study shows demand redistributing, not shrinking: travel is the resilient vertical.
July 2026
The state of play
The industry, in sourced numbers
The state of travel 2026: 40 numbers that explain the year
307 million Q1 arrivals, a record 847 million across the OECD, $1.37 trillion in US spend, and the great bifurcation: fewer travellers, spending more each.
July 2026
What luxury travellers want in 2026: the behaviour shift in 30 statistics
Pet bookings up 260%, luxury rail the most-wanted stay, milestone trips filling calendars, and wellness compounding towards $2.1 trillion.
July 2026
The price of a click: what advertisers pay for luxury travel searchers
Advertising’s honest metric: arranger clicks cost roughly three times product clicks, and “luxury travel company” now out-searches “luxury travel” itself.
July 2026
The demand ledger
Where the world wants to go
The most-searched travel destinations of 2026
34 cities ranked by 44,089,019 search impressions. Osaka leads, Japan takes four of the top eight, and 22 European cities finish within 8% of each other.
July 2026
The Travel Demand Index 2026
What 62 million search impressions reveal: the Japan surge, the solo-travel spike, and the budget-and-luxury split-screen consumer.
July 2026
The luxury travel search map 2026
Luxury co-search trends across seven markets: agents declining everywhere, companies rising, and India up 300% in British searches.
July 2026
The category reports
One market, measured, per report
The travel insurance search report
The grey boom: over-70s cover up 1,400%, US comparison searches up 450%, and Britain’s comparison sites losing the comparison.
July 2026
The travel credit card search report
Amex research attention down 80%, “best credit cards 2026” breaking out, and a market splitting into premium rewards and credit repair.
July 2026
eSIM search trends 2026
The education boom in one dataset: “what is esim” up 90%, challenger brands breaking out, and one provider alone in the US head terms.
July 2026
Citing this research
Journalists, researchers and publishers are welcome to quote or reproduce any figure from an Uncompromised Data report with attribution to Uncompromised Travel and a link to the report it comes from. Proprietary findings, including the citation monitor, the ChatGPT readings and the grounding-query analysis, are published here first and exist nowhere else.
For methodology questions or the underlying numbers behind any chart, write to richard@uncompromised.travel. Companies that want to know where they stand in these answers can start on the For Brands page. The full editorial archive lives under Travel Intelligence.