Beyond Tokenomics: The Threat of the Invisible Hotel

Earlier this month at Skift's AI Summit, Hilton's Michael Leidinger took the stage and introduced the industry to a new term: tokenomics. He was talking about the cost of running AI, the price hotel brands pay when their own agents and the outside agents crawling their sites start racking up a bill. It landed because it named something every finance team is already feeling. Skift, through Rafat Ali and Adriana Lee, framed it well: travel faces a two-sided squeeze, where AI agents search sites at volumes no human ever could, breaking the old look-to-book math, while the AI you run in-house keeps costing more even as per-token prices fall.

That is a real problem. But it is only half the P&L.

Here is the part the cost story leaves out. All of that spending, all of that tokenomics, assumes the AI is searching your site in the first place. The more consequential question is not what an AI query costs you. It is whether the AI names you at all. And the latest data says that the answer is concentrating fast, among a small group of winners, chosen by sources that most brands do not yet manage.

I have spent 20 years in luxury hospitality sales and marketing. I have watched this movie before. It was called the OTA. We spent a decade fighting over ranking on someone else's page, paying for position, and learning the algorithm a beat too late. AI discovery is the same fight moving upstream, from the booking page to the layer where the recommendation is actually made. If you are only optimizing the meter, you are solving last year's problem.

The numbers make the case. Yext analyzed 6.8 million hospitality citations and found that major AI platforms do not agree with one another. Gemini leans heavily on brand-owned sites. ChatGPT and Perplexity lean on aggregators and third parties. There is no single switch you flip to be visible everywhere, because each model trusts a different set of sources. Then there is 5W's AI Visibility Index, which found citation share concentrating hard, the top three brands taking the lion's share in several categories. The kicker: loyalty-program size did not predict who got cited. The biggest rewards program does not buy you the recommendation. And Hotelrank found that roughly 65% of hotel queries trigger a live web search, which means the answer is being assembled in real time, from live sources, on most of the questions your future guests are asking.

Put those together, and the picture is clear. Being big is not the same as being visible. The recommendation is contestable on most queries. And the sources that decide it are editorial and third-party, the ones sitting upstream of your booking engine, the ones most brand teams are not managing yet.

This is why visibility is the new distribution. The brands that win the next decade will not be the ones who shave a few cents off the per-query bill. They will be the ones planted in the sources these models actually read and cite. Cost discipline without a citation strategy is polishing a meter nobody will ever look at.

Now, the fair pushback. Cost is the more urgent problem because it is measurable today, and citation share is volatile, unaudited, and swings with every prompt and model update. All true. But volatility cuts both ways. Volatile markets are exactly where early, disciplined players build a compounding lead. GEO services will commoditize eventually, the way SEO did. The brands building structured authority now, before the field crowds in, are the ones who will be named by default when the dust settles. Ask any endurance athlete: the work you put in early, when it is quiet and nobody is watching, is what shows up at mile 80.

So what does this mean if you run a hotel or a portfolio? A few things. Citation share becomes a board-level number, sitting next to RevPAR and your direct-booking mix, because it feeds both. Earned media and structured data move upstream, from a nice-to-have at the bottom of the funnel to the thing that decides whether you enter the funnel at all. And loyalty scale stops being your visibility proxy. Being the biggest name in the room does not help if the AI does not say your name.

Watch two things over the next year. First, whether someone publishes a recurring, quarterly citation-share index that becomes the industry's scoreboard, the STR of AI visibility. Second, watch which major chain is first to name a Head of AI Visibility. When that title shows up on a real org chart, you will know the recommendation layer has gone from a marketing curiosity to a distribution mandate.

Skift got the cost story right. Every operator should be watching their tokenomics. But do not let the meter distract you from the front door. The next distribution war is not about what the query costs. It is about whether you get named. Visibility is the new distribution, and the brands treating it that way right now are the ones who will still be in the answer when the rest of the field finally looks up.