I’ve sat through three vendor briefings in the past two months where the same slide appeared. A widely cited McKinsey analysis put a number on it: AI personalisation and dynamic pricing could boost hotel revenue 3 to 10 percent annually — a figure that has become the industry’s favourite justification for accelerating AI deployment. Every major conference in 2026 carries some version of this headline — that AI will finally deliver luxury-grade, individually tailored experiences at every price tier. Over 60 percent of the world’s top 100 luxury hotel brands had launched AI personalisation pilots by 2026, according to market research tracking the AI-powered luxury travel segment.
The question nobody is asking: if every hotel deploys the same AI personalisation stack from the same handful of vendors, what exactly is being personalised?
What AI Hotel Personalisation Is Actually Proposing
Today’s hotel AI personalisation works by connecting loyalty history, booking behaviour, and property management data into preference profiles. It knows you booked a high floor three times in a row. It knows you ordered room service at 11pm. It knows you checked the gym schedule on arrival. These are functional attributes — observable, recordable, rankable.
What it cannot do is read the sentence in your last review that says “the view from room 812 stopped me cold.” It cannot score the sense of arrival at a lobby designed by a world-class architect. It cannot detect that the reason you rebook a specific San Francisco hotel has nothing to do with rate or location — it is a quality of place you would struggle to name but would instantly recognise. I’ve been trying to articulate this for years in client hotel programme reviews, and I still can’t do it cleanly.
This is not a gap that more data closes automatically. It is a structural limitation: AI ranks what it can measure.
Why Corporate Travel Managers Need to Ask a Specific Question
Corporate travel has historically treated hotels as functional infrastructure — rate, location, fitness centre. That calculation is shifting. Eighty-three percent of business travellers took a bleisure trip in the past year, extending stays that would otherwise average two to three days into four-to-seven-day trips. The bleisure market reached an estimated $472 billion in 2025, according to Research and Markets. Travellers are increasingly asking for an experience worth staying for.
More directly: when two hotels in a preferred programme sit at similar negotiated rates, experiential quality becomes the tiebreaker that drives compliance. The question worth adding to your next sourcing review is not only whether your AI tool enforces rate caps — it is whether the preferred properties you selected for their character and employee satisfaction will still feel the same after their AI stacks are fully deployed and their human service layers have been partially automated. That shift is already happening at hotels in your programme.
Where Experience Actually Lives in the Booking Stack
Luxury signals exist in the data — just not where AI booking tools look for them. User reviews are dense with experiential vocabulary: “breathtaking views,” “the room was opulent,” “a sense of arrival you cannot manufacture.” Tools like Revinate and TrustYou already parse this sentiment at scale, extracting quality signals from thousands of reviews. The training data exists.
The problem is that this experiential intelligence lives in the reputation layer, not the booking layer. Global booking systems carry amenity flags — pool, restaurant, fitness centre — not “harbour view that justifies the premium.” The infrastructure that shortlists hotels for most corporate travellers was built to compare functions, not experiences. AI booking agents shortlist from that layer. The luxury signal never reaches the ranking algorithm.
LVMH eventually concluded that a shared multi-brand platform — eLuxury.com, which ran from 2000 to 2009 — could not preserve the distinct experience each brand required. They shifted to individual brand-owned channels. Hotels have not yet made that equivalent design choice in the AI booking stack.
The Homogenisation Problem AI Is About to Make Worse
The risk has two layers. The first involves execution: hotels that deploy AI without encoding their brand experience risk producing outputs functionally indistinguishable from competitors running the same stack.
The second layer is structural and harder to solve. Most hotels — across all tiers — are sourcing their AI personalisation tools from the same short list of vendors. The problem is not how any one hotel configures its system. The problem is that every property on the same platform shares the same underlying model, trained on the same industry-wide data. A boutique design hotel and a mid-scale business property can encode entirely different brand values into the same vendor’s system — but they are still running the same algorithmic engine. Distinctiveness as an input does not guarantee distinctiveness as an output.
The commercial stakes are significant. Boutique hotels command a 34 percent ADR premium, averaging $258 per night in 2025, according to Highland Group — with luxury-tier independents commanding significantly more. U.S. luxury hotel RevPAR grew approximately 3 percent in early 2025 while economy chains contracted by over 4 percent — the widest performance gap between chain scales in years. That premium is the market’s demonstrated willingness to pay for genuine differentiation — exactly what vendor-concentrated AI will quietly erode.
The Bigger Risk: Neither Side Can Afford Homogenisation
AI will not flatten luxury hotel experiences — not because the technology is incapable, but because both sides of the transaction have powerful incentives to resist it. Travellers extending business trips into bleisure are actively seeking experiences worth staying for. Hotels earning a 34 percent rate premium are financially dependent on the differentiation that produces it.
What will force the evolution is commercial pressure and training data. The review vocabulary that teaches AI to recognise “breathtaking view” already exists at scale. The missing step is intentional: the industry needs to redesign how reviews capture sensory and experiential language in machine-readable form — and travel managers need to demand that their booking AI shortlists on the full picture of quality, not only the fraction it currently measures.
The luxury experience is not beyond AI’s reach. It is beyond AI’s current training data. That is a solvable problem — and one I think this industry will be forced to solve faster than most vendors are currently planning for.
Rajeev Goswami is CEO of WWStay and a member of the GBTA Technology Committee.
Sources
- AI personalisation and dynamic pricing in hospitality — McKinsey analysis
- AI-Powered Hyper-Personalized Luxury Travel Experience Market — MarketIntelo
- 50 Bleisure Travel Statistics for 2026 — HotelTechReport
- 2026 Bleisure Travel Trends — Engine
- Bleisure Travel Market 2025 — Research and Markets via SkilTravel
- Luxury independent boutique hotels hit top performance in 2025 — Highland Group / Boutique Hotel News
- Top hotel industry statistics 2026 — SiteMinder
- LVMH eLuxury.com — South China Morning Post
- Guest Sentiment Analysis — TrustYou
- AI in Hospitality: The 2025 Reality and the 2026 Horizon — Hospitality Upgrade
- Benchmarking Beyond Rates: Hotel Performance Metrics — ReadyBid
Leave a Reply