How to Get Your Hotel Onto AI’s Shortlist

The Place Firenze transformed itself from JK Place into one of Florence’s most distinctive boutique hotels. So why was AI still telling yesterday’s story? Lucy Clifton explores where ChatGPT hotel recommendations actually come from, and what happened when Spotlight and Make Lemonade set out to discover what AI really knew about the hotel.

Piazza Santa Maria Novella wakes early. The first commuters cross from Florence’s main station, while tourists pause to photograph the green serpentine and white Carrara marble of Leon Battista Alberti’s basilica and café owners set up tables. Within minutes, the square has become what it has been for centuries: one of Florence’s great meeting places, once filled with friars, pilgrims and merchants, today with backpackers, hotel guests, street performers, artists and students, all briefly occupying the same stage.

Looking out across it all is The Place Firenze, Italy’s original townhouse hotel, which reopened in spring 2026 following the final stage of its transformation from JK Place Firenze. Staying here feels like being the guest of an exceptionally cultured friend. Guests check in at an antique desk in a library lined with books on art, architecture and photography. Upstairs, rooms redesigned by Florentine architect Luigi Fragola look across terracotta rooftops towards the city’s domes and campaniles. And, at the heart of the hotel is general manager Claudio Meli, who has spent decades building relationships across the city and connects guests with Florentine artisans, chefs, restorers and makers. It is a hotel with a strong sense of its own identity, shaped as much by Florence itself as by the people who run it.

Editorial acclaim, AI silence

One of the first places I look when researching hotels is The Telegraph, which under Hotel Editor Rachel Cranshaw has one of the strongest editorial hotel databases in the world with more than 10,000 reviews. It may not have the glamour of Condé Nast Traveller or the global reach of Travel + Leisure, but in terms of depth, consistency and trust, it is exceptional. So I Google ‘The Telegraph Best Hotels in Florence’ (in incognito mode) and at the top of the list is The Place. With a score of 9/10, their Florence expert, Nicky Swallow, describes it as, “one of the best city boutique hotels in Italy – the immaculate interiors, cocooning atmosphere and exceptional service make this the perfect Florentine pied-à-terre.”

Next, I ask ChatGPT a broad question: “I want to visit Florence. What are the city’s best hotels?” It returns a familiar list: Four Seasons, The St Regis, Hotel Savoy, Hotel Lungarno, Villa San Michele, Hotel La Gemma. The Place Firenze does not appear. The editorial endorsements are there. ChatGPT hotel recommendations do not necessarily follow.

One conversation proves nothing, of course. But it neatly illustrates a question that has become central to Spotlight’s mission: why do some excellent hotels become part of AI’s mental shortlist while others do not?

Why PR and AI intelligence had to meet

It marks a profound change for luxury travel and a shift that brought my agency, Spotlight, together with Make Lemonade, the AI consultancy founded by Maria Sze and Sara Lemos. Spend five minutes with either of them and you will quickly realise they are not interested in making AI sound more complicated than it needs to be. Instead, their work is about translating a fast-moving technology into something businesses can actually use. We quickly realised that neither PR nor AI intelligence could solve the problem alone. Understanding what AI has learnt about a brand is one challenge; changing AI’s understanding requires editorial authority, trusted relationships and strategic communications.

That collaboration led to The Full Story, a new service developed by Spotlight and Make Lemonade to help luxury travel and lifestyle brands understand and improve how they are represented by AI. At its heart is Signal Noir, a proprietary methodology that reveals the hidden signals shaping AI’s view of a brand.

We called it Signal Noir because AI is not reading brands in the way people do. It is piecing together thousands of small signals that most businesses never even realise they are sending. On their own they seem insignificant. Together, they become the story AI tells.

Maria Sze, Founder, Make Lemonade

Rather than focusing on algorithms or rankings, Signal Noir asks a deceptively simple question. As Maria puts it: “When someone asks AI for a recommendation, what does AI actually know about your brand?”

Putting Signal Noir to the test

As one of Spotlight’s newest clients, The Place Firenze became one of the first hotels to put Signal Noir to the test. Leading the audit is Make Lemonade’s strategist, Sara Lemos, whose career spans two decades advising brands including Michelin, Swarovski, Emirates and Porsche. For The Place Firenze, Sara tested 30 luxury travel queries across three traveller personas, two markets and four AI platforms, examining not just whether the hotel appeared, but what story AI was actually telling about it.

Ask AI about The Place Firenze, and it paints a broadly accurate picture. It strongly associates the hotel with its transformation from JK Place Firenze, recognising it as an independent boutique hotel overlooking Piazza Santa Maria Novella. But beyond those fundamentals, the picture becomes surprisingly thin. The hotel’s dining and rooftop views barely register. It can describe the hotel; it has not yet built a complete picture. What ChatGPT, Gemini or Perplexity return is not simply a reflection of a hotel’s quality. It is a reflection of everything the internet has taught them about it. The facts are there; the narrative is not.

While The Place Firenze has spent the past few years reinventing itself with a new name, a new design language and a philosophy centred on connecting guests with contemporary Florence, AI still largely remembers the hotel it used to be.

Why ChatGPT hotel recommendations change by market

The audit’s first surprise was that there was not a single AI answer. When Signal Noir tested identical prompts using US and UK traveller profiles, The Place appeared far more often in recommendations for the US than the UK, revealing that AI’s view of a hotel can change depending on who is asking. “People often assume there’s one definitive AI answer,” Sara tells me. “There isn’t. Recommendations change depending on the market, the traveller profile and the context of the question.”

In the US, AI’s understanding of The Place was already being reinforced by booking and review platforms such as TripAdvisor. In the UK, however, Sara identified a different opportunity. No single hotel dominated AI’s recommendations, suggesting there was still plenty of room for stronger editorial signals to shape the conversation.

Who AI was learning from

The second discovery was who AI was learning from. AI was not making random choices or favouring bigger hotel groups. It was consistently drawing on the sources it appeared to trust most. Traditional titles such as The Telegraph, CN Traveller, The Times, Harper’s Bazaar, Forbes, Michelin Guide and Mr & Mrs Smith sat alongside specialist travel sites including The Hotel Guru, A Small World, The Florence Insider and Ciao Bambino. These were the publications and platforms that consistently appeared in AI responses when recommending competitors over The Place. They were not simply nice places to earn coverage; they had become part of AI’s education.

As a PR, for me this means the question is not “How many pieces of coverage did we get?” but “What story did those pieces collectively teach AI?”

We used to think about editorial as something that persuaded readers. Now it also teaches AI. That’s a fundamental shift in how brand visibility works.

Sara Lemos, Strategist, Make Lemonade

For Caroline Bondy, Director at Spotlight Communications, who leads The Place Firenze account, that insight has already influenced how our team thinks about media strategy.

We’ve always valued coverage in top-tier publications, but Signal Noir showed us that highly relevant editorial can play an equally important role. A mention of The Place Firenze in a specialist feature on rooftop bars, for example, may have a far greater influence on how AI understands the hotel than its readership alone would suggest. It’s no longer just about where a story appears, but what it teaches AI about the brand.

Caroline Bondy, Director, Spotlight Communications

The gaps in the record

The audit also exposed some surprisingly simple gaps. While JK Place Firenze had established a footprint across sources such as Wikipedia, The Place’s new identity was less consistently represented.

“People often assume AI knows everything about their brand,” Sara says. “But it can only work with the information that’s available to it. Even small gaps in how a hotel is represented across trusted sources, or on its own website, can affect the evidence AI has when deciding what to recommend.”

Finally, there was the brand narrative itself. Signal Noir found that AI already had a strong understanding of The Place as an independent boutique hotel with character, authenticity and a successful rebrand from JK Place. But its understanding was incomplete. Themes such as dining, rooftop views, celebrations and family stays were not being reinforced consistently enough across the hotel’s own channels and trusted editorial sources.

AI does not infer a brand’s positioning from a single page or article. It learns through repetition.

Unless the same themes appear consistently across a website, editorial coverage and third-party sources, AI is less likely to associate them confidently with the brand.

How a hotel teaches AI its story

So how does a hotel teach AI its story? “The answer isn’t gaming algorithms or chasing keywords,” Sara explains. “It’s making sure the same narrative appears consistently wherever AI is likely to learn.” For The Place, the conclusion was simple: AI had not overlooked the hotel; it simply had stronger evidence for recommending others.

Back in Piazza Santa Maria Novella, the square has not changed, but how people find it has. For many of the travellers wheeling suitcases across the square, the journey probably began with a conversation on ChatGPT. The hotels that thrive from here will not necessarily be those shouting the loudest. They will be the ones whose stories are told consistently enough, by journalists, guests, partners and the brands themselves, that AI learns to tell them too.