We asked ChatGPT to recommend a private, all-inclusive resort in the Maldives. Top of the list was Spotlight’s client Kudadoo. Then we asked which independent journalism supported its choice. Lucy Clifton reports on the results.
I want a very private Maldives resort for a special occasion, preferably all-inclusive rather than having to think about extras. What are your top five recommendations?
So went my (incognito) request to ChatGPT on 1 September 2026. I didn’t mention a hotel. I didn’t mention Spotlight. And I didn’t mention any of the publications that have written about our clients.
ChatGPT put Kudadoo Maldives Private Island at number one. The next four were Soneva Jani, Velaa, Hurawalhi and Milaidhoo. Two declarations before I go any further: Hurawalhi is also a Spotlight client, and Kudadoo and Hurawalhi are sister properties under the same ownership. I told ChatGPT none of that. Its summary read:
I’d book Kudadoo. It most closely matches your specific brief: very private, special occasion, luxury, and I don’t want to think about the bill. Its tiny scale and genuinely comprehensive all-inclusive concept make it unusually compelling.
ChatGPT
As Kudadoo’s PR, this was obviously the result I was hoping for, and it was a reply that interested me. So I asked a follow-up question.
Why did you recommend Kudadoo specifically? What sources or independent evidence are you relying on? Please give me the publications or websites that most influenced that recommendation.
ChatGPT pointed first to a Condé Nast Traveler review by Lisa Grainger, describing this as probably the single most relevant independent source for the recommendation. It also cited Forbes Travel Guide and Robb Report Singapore among the editorial sources supporting its assessment of the resort’s privacy, luxury and all-inclusive credentials.
So I pushed it further. What did it consider independent evidence, and what didn’t it?
Its answer drew a striking distinction. Kudadoo’s own website, it said, was useful for verifying what the resort currently offers, but not as evidence that the resort is good.
Kudadoo could tell AI what it was. Independent journalism helped AI decide whether to believe it.
PR’s second audience
For decades, the value of earned media, meaning independent coverage you haven’t paid to place, from newspaper reviews to magazine features, was relatively easy to understand. A feature in Condé Nast Traveler, The Times or Forbes Travel Guide put a hotel in front of the right people and, crucially, gave someone other than the hotel permission to say it was exceptional.
That hasn’t changed. What has changed is who, or what, might be reading it.
As travellers increasingly ask ChatGPT, Claude, Gemini and Perplexity where to go and where to stay, those systems have to construct their answers from evidence. A hotel’s own website is part of that picture. So are booking platforms, review sites, databases and other material across the web. But independent editorial coverage can do something a hotel’s own marketing cannot: provide third-party corroboration.
Our Kudadoo experiment offers a neat example. ChatGPT didn’t simply find the resort. It used independent journalism to help explain why Kudadoo fitted the brief.
PR has always been in the business of building reputation with people. Increasingly, the evidence it leaves behind may also help shape how machines describe, compare and recommend a brand.
What the research tells us
Kudadoo is an illustration, rather than proof, of how AI recommendations work. But it reflects a much broader pattern.
In research carried out by Spotlight Communications and Make Lemonade, 30 leading luxury travel and lifestyle publications were tracked across four major AI platforms, generating more than 1,400 citations. The results showed that AI systems are already relying on travel journalism when answering questions about where to go and where to stay, although which publications they favour varies considerably from platform to platform.
That matters because AI is doing more than finding information. Increasingly, it is assembling an answer from multiple sources: comparing claims, looking for specificity and deciding which sources are useful enough to cite or draw upon.
Independent coverage sits in an interesting position within that mix. A hotel can describe itself as secluded, sustainable, exceptional or one of the best in the Maldives. A journalist who has actually been there can independently report the things that make those claims credible.
And there is evidence that the way information is supported affects whether it surfaces in generative search. Researchers from Princeton University, Georgia Tech, IIT Delhi and the Allen Institute for AI found that techniques including adding credible citations, quotations and statistics could increase the visibility of source material in generative-engine responses, in some cases by up to 40 per cent.
That doesn’t mean a glowing review in Travel + Leisure automatically propels a hotel to the top of ChatGPT’s recommendations. AI systems are more complicated, and considerably less transparent, than that.
But it does suggest something PR professionals have understood instinctively for decades: who says something matters as well as what is said.
Not all coverage is equal
The Kudadoo example also reveals something else. ChatGPT didn’t recommend the resort simply because Condé Nast Traveler had written about it. It could explain which details in that coverage answered our particular question.
We had asked for privacy. The review described a tiny, secluded island with just 15 residences, where it was possible to go through a day without seeing another guest. I wanted all-inclusive. The journalism ChatGPT cited in support of its answer explained how Kudadoo’s proposition went far beyond meals and drinks. ChatGPT replied that the resort’s “Anything, Anytime, Anywhere” philosophy was unusually comprehensive, with dining, premium drinks, spa, diving, watersports and excursions all capable of being included rather than constantly added to the bill.
I said it was for a special occasion. ChatGPT explained:
Forbes Travel Guide describes Kudadoo as an ultra-exclusive escape with just 15 expansive overwater Ocean Residences. I put weight on this because Forbes Travel Guide is useful for answering a slightly different question from a normal guest-review site: is this actually operating at the level of luxury implied by the price?
ChatGPT
In other words, the useful part wasn’t simply the masthead. It was the specificity. And, intriguingly, not all the evidence ChatGPT valued was glowing.
It singled out a first-hand Robb Report Singapore review by Charmaine Tai dating from 2020. While it was overwhelmingly positive, Charmaine did criticise aspects of the arrival and transfer experience. ChatGPT told us that this actually made it more confident in the recommendation because the article wasn’t uniformly promotional.
That’s an important distinction.
Independent journalism doesn’t derive its authority from saying nice things about a hotel. It derives it, at least in part, from being independent.
But professional journalism wasn’t the only independent evidence ChatGPT considered. It also pointed to recent TripAdvisor reviews as supporting evidence that Kudadoo was still delivering what the journalists had described. One guest called its all-inclusive proposition the real deal; another described the privacy as exceptional.
That raises an interesting question about where independent evidence begins and ends. One travel journalist we spoke to told us that, after reviewing hotels professionally for newspapers and magazines, she has sometimes subsequently been asked by the hotel to post a review on TripAdvisor too. Some of Kudadoo’s own TripAdvisor reviews are similarly marked as having been collected in partnership with the hotel.
That doesn’t make them paid-for endorsements, nor does it make a TripAdvisor review equivalent to a reported piece in Condé Nast Traveler. Indeed, when we asked ChatGPT to distinguish between its sources, it gave greater weight to professional editorial coverage. But guest reviews add another layer of corroboration: the hotel says what it promises; journalists independently report what they experienced; guests report whether the promise was delivered.
Together, they create something much harder for a brand to manufacture: corroboration.
That has implications for the stories hotels choose to tell. “Barefoot luxury”, “unforgettable experiences” and “a unique sense of place” may sound seductive in a press release, but they give an AI system very little concrete information with which to distinguish one hotel from another.
A journalist reporting that a resort has 15 residences, that guests can dine wherever and whenever they choose, or that a particular experience is included in the nightly rate is leaving behind something much more useful: specific facts and first-hand observations that can be retrieved, compared and attributed.
For hotel PR, then, the opportunity isn’t simply to generate more coverage, or even universally positive coverage. It is to give good journalists something distinctive and substantive to report.
The best stories have always done that. AI just gives us another reason why it matters.
So does PR need to change?
Yes, but perhaps not in the way you might think. The answer isn’t to start writing press releases for ChatGPT, or asking journalists to shoehorn particular phrases into their reviews. If anything, our Kudadoo experiment points in the opposite direction.
The things that made its coverage useful to AI were the things that make journalism useful to people: independent reporting, first-hand experience, specific details and credible sources. A respected travel writer actually visiting a hotel can establish things that no amount of carefully optimised brand copy can quite replicate.
What does change is how PR teams think about the life of that coverage. A newspaper review or magazine feature was once primarily judged by the people who might read it: its audience, influence and ability to inspire a booking. Those things still matter. But published journalism can now become part of the body of evidence AI systems encounter when they are asked to compare hotels and make recommendations.
That makes the job both broader and more exacting. Hotels need a clear, consistent story across their own digital presence, but they also need independent voices supplying the detail, context and corroboration that a brand cannot provide for itself. It is why AI reputation and search visibility now sits alongside brand PR rather than apart from it.
It also means resisting the temptation to measure PR simply by volume. Ten pieces of generic coverage repeating the same press-release language may be less useful than one authoritative, detailed review by a journalist who has actually experienced the property.
The new value of a great review
Which brings us back to Kudadoo.
When I asked ChatGPT where to stay in the Maldives, the resort didn’t appear because I asked about Kudadoo. I hadn’t. It appeared because I described the kind of hotel I wanted, private, special and genuinely all-inclusive, and ChatGPT decided Kudadoo fitted the brief.
When I asked it to show its working, independent journalism was part of the answer.
That doesn’t mean a Condé Nast Traveler review guarantees a place on an AI shortlist, any more than it ever guarantees a booking. But it does change how we should think about the value of that review.
A great piece of coverage can still inspire the person who reads it over breakfast on a Sunday morning. It can still build reputation, create desire and put a hotel on someone’s mental wish list.
But increasingly, it may also be encountered by an AI system months or years later, when a completely different traveller asks a completely different question.
For PR, that’s a significant shift. Earned media isn’t becoming less important as travellers move from search engines to answer engines. It may be acquiring another job entirely.
Kudadoo didn’t tell us it was the right hotel. ChatGPT did. And when we asked why, it pointed us towards the journalists.
Invisible or Influential? How Leading Travel and Luxury Publications Are Adapting to AI tracked 30 leading luxury travel and lifestyle publications across major AI platforms, alongside interviews with senior editors. Download the full research.
