What AI Made of “The Bully Bought the Cafeteria”

Hotels spent years trying to win more direct bookings. In “The Bully Bought the Cafeteria,” Americas Great Resorts asks who shapes the traveler’s choice before the booking. We asked Google AI, ChatGPT, and Gemini for their thoughts. Their answers differ on how much influence online travel agencies (OTAs) have, where it comes from, and what it means for hotels.

Series · AI Reviews of AGR Articles
Original article · The Bully Bought the Cafeteria (May 19, 2026)
Captures and AGR review · November 4, 2026. The same prompt was run in each system in a private Firefox window while logged out.

Exact prompt
Tell me what your thoughts are on this article: https://www.americasgreatresorts.net/bully-bought-the-cafeteria/

AGR’s reading of the responses

The sandwich stands for the direct-booking channels hotels built to reduce OTA dependence. The new room stands for AI-mediated discovery. A hotel can own the channel where a traveler books while other sources shape how that traveler first encounters the property.

Google AI. Google AI read the metaphor through the lens of training data. The article invites that reading by describing Expedia, Booking.com, and TripAdvisor as teaching the room about hotels. Google AI went further. It attributed a model’s preference for OTA descriptions to the scale of OTA content. It listed “Periscope” as an AI search engine without identifying the product and expanded KFO as “Knowledge Engine Optimization.” AGR’s term is Knowledge Formation Optimization.

ChatGPT. ChatGPT challenged the idea that Expedia controls AI discovery. It distinguished being recommended, being accurately described, and winning the booking. It cited research from Phocuswright, Cornell, and Mower on how travelers use AI and what they do with its suggestions. Which sources shape a particular hotel recommendation is a separate question. ChatGPT’s source labels are reproduced as supplied.

Gemini. Gemini read the article primarily as a story about hotels winning tactical battles while OTAs gained influence over distribution. It described negotiations over commissions, rate parity, and contracts that do not appear in the article. The article’s hotels pack their own lunch by building direct-booking channels. Gemini made its connection to AI through AGR’s broader work and left the new room, the teacher, and the report card out of its reading.

Google AI and ChatGPT also noted AGR’s commercial interest. AGR provides AI visibility services and reports for hotels.

Google AI Response

This article presents a sharp, metaphorical critique of how the hotel industry risks falling behind in the age of artificial intelligence. The core argument is that while hotels spent a decade trying to bypass online travel agencies (OTAs) like Expedia through direct booking campaigns, OTAs have already dominated the next digital frontier: training AI models on hotel data.

Here is a breakdown of the article’s key themes, why its metaphor works, and the underlying strategic reality for modern marketing.

🥪 The Metaphor Breakdown: “Packing a Lunch”

The author uses a clever playground analogy to describe a massive corporate shift:

• The Bully: Online Travel Agencies (OTAs) like Expedia and Booking.com, which traditionally command high commission fees from hotels.
• The Packed Lunch: Direct-to-consumer booking strategies (loyalty programs, website optimization) that hotels built to bypass OTA commissions.
• The Cafeteria: The entire digital ecosystem. The article argues that celebrating direct booking success is short-sighted because the ecosystem itself has shifted to AI-driven discovery.

🤖 The Reality of “The New Teacher”

The article highlights a critical technical reality of large language models (LLMs) and AI search engines:

• Data Bias in AI Training: AI search engines (like ChatGPT, Periscope, Gemini, or specialized travel bots) learn from the most dominant data footprints on the web.
• The Footprint Disparity: OTAs have spent millions of pages and decades structuring detailed reviews, metadata, and descriptions of hotels. Because their data footprint is massive, AI models default to the OTA’s perspective and description of a property rather than the hotel’s own narrative.
• The “Report Card” Effect: If an AI engine recommends a hotel based on what Expedia says about it, the hotel loses control over its own positioning, branding, and ultimately, the booking pipeline.

⚖️ Critical Analysis: Where the Article Lands

• What it gets right: The author is entirely correct about the shift toward Generative Engine Optimization (GEO) or LLM Optimization. Relying strictly on traditional SEO or standard direct-mail lists is no longer enough when consumers use AI as a personalized concierge to filter accommodation options.
• The underlying sales pitch: It is worth noting the context of where this article is published. The site explicitly sells “KFO” (Knowledge Engine Optimization) and “AI Visibility Reports” for hotels. While the warning is grounded in real tech trends, the framing is designed to create a sense of urgency (FOMO) to drive demand for their specific visibility services.

🚀 The Next Strategic Steps

For any hospitality brand reading this piece, the practical takeaways are clear:

• Claim your data footprint: Structuring website data (using Schema markup) so AI crawlers can accurately read prices, amenities, and unique selling points directly from the source.
• Audit AI visibility: Testing how major LLMs respond to prompts like “What is the best luxury boutique hotel in [City] for a couple?” to see if your property is even in the conversation, and why.
• Maintain the “sandwich”: Direct booking channels are still highly profitable and necessary; the goal shouldn’t be to abandon them, but to ensure AI engines know how to direct users to them.

Are you looking at this article from a hospitality marketing perspective, or are you interested in how AI optimization (GEO/SEO) works across industries more generally? Let me know so we can dive into specific tactics.

ChatGPT Response

My initial reaction: I think this article makes an important point, uses an excellent metaphor, and identifies a genuine strategic risk for the hotel industry. But it also deliberately exaggerates the situation to make its point.

I would take the underlying argument seriously, even if I wouldn’t accept every assertion at face value.

The article, The Bully Bought the Cafeteria, is really about a shift in who controls the relationship between hotels and travelers. And I think that’s a much bigger issue than simply whether hotels are getting enough direct bookings.

1. The central argument is compelling

The author uses Expedia as the bully and the cafeteria as the distribution system.

The hotel industry’s response to online travel agencies (OTAs) has largely been to develop ways to encourage direct bookings:

• Build better hotel websites.
• Offer loyalty programs.
• Promote direct-booking benefits.
• Invest in SEO and paid search.
• Develop email databases and customer relationships.

The article argues that hotels have been concentrating on controlling their own little piece of the distribution system while Expedia and other OTAs have been building influence across the entire ecosystem.

Then comes the next development: AI-powered travel discovery.

The author’s argument is that travelers are increasingly asking AI where they should stay, and AI is drawing heavily on information published by Expedia, Booking.com, TripAdvisor and other established platforms.

The particularly important observation is this:

If AI learns about your hotel primarily through someone else’s description of it, someone else is effectively influencing how your hotel is presented to prospective customers.

That’s a legitimate concern.

Americas Great Resorts

2. The timing is supported by actual industry data

This is where I think the article becomes particularly interesting.

The shift toward AI-assisted travel discovery is no longer theoretical.

A March 2026 study by Phocuswright reported that:

• 56% of active U.S. travelers had used AI for planning, booking, or assistance during a trip in the previous 12 months.
• That figure had risen from 33% in the first half of 2025.
• AI platforms were approaching traditional search engines in travel product research.

Phocuswright
+1

A separate 2026 study from Cornell University examined 1,029 U.S. travelers and found that AI is particularly useful for travel discovery and evaluating hotel recommendations.

Cornell eCommons

So the author’s fundamental warning is justified: hotels cannot assume that Google and traditional search will remain the primary gateways to their properties.

3. Where I disagree with the article

I think the article makes three arguments that deserve more scrutiny.

First: Expedia does not own the AI cafeteria.

The analogy is powerful, but it oversimplifies how AI systems acquire and use information.

AI platforms draw on many sources, including:

• Hotel websites and official property information.
• Guest reviews and travel publications.
• OTA listings.
• Destination marketing organizations.
• News articles, forums and other publicly available information.

Expedia has influence, but it doesn’t control the entire information ecosystem.

In fact, the opportunity for independent hotels and hotel brands is that they can develop their own authoritative digital presence.

Second: Being visible in AI recommendations isn’t the same as winning the booking.

This is an important distinction.

A hotel might be recommended by ChatGPT, but the traveler may subsequently:

• Visit the hotel’s website.
• Compare prices on Expedia.
• Read TripAdvisor reviews.
• Look at photographs on Google.
• Book through a completely different channel.

Research published by Mower in September 2026 found that only 5% of surveyed AI-using travelers booked exactly what AI suggested. Many went on to verify recommendations through Google and official websites.

MOWER

That means AI visibility is important, but it isn’t the whole commercial equation.

Third: The article doesn’t sufficiently distinguish between visibility and accuracy.

I would argue that the real objective shouldn’t simply be getting AI to mention a hotel.

It should be getting AI to understand the hotel correctly.

A luxury resort, for example, needs to be understood in terms of:

• Its actual positioning and target clientele.
• Its location and surrounding experiences.
• Its distinctive facilities and amenities.
• Its service philosophy.
• Its strengths relative to competing properties.

Being recommended for the wrong reasons can be almost as problematic as not being recommended at all.

4. What I think hotel executives should take away

I would frame the strategic opportunity somewhat differently from the article.

The next competitive advantage isn’t simply direct booking versus OTA booking.

It’s control over hotel discovery, representation, reputation and demand generation.

I would prioritize five things:

1. Own your information. Your official website should be the most comprehensive, accurate and authoritative source about your property.
2. Manage your external footprint. Understand how Expedia, Booking.com, TripAdvisor, Google, travel publications and review platforms describe your hotel.
3. Develop AI visibility. Regularly evaluate whether ChatGPT, Gemini and other AI platforms recommend your property for relevant traveler searches.
4. Measure the commercial impact. Track AI-referred traffic, brand searches, qualified inquiries, website conversions and direct bookings. Don’t confuse AI mentions with revenue.
5. Preserve the guest relationship. Regardless of where discovery happens, the hotel should have a strategy for converting that interest into a direct relationship.

5. One thing I think readers should recognize: there’s a commercial motive

I would also look at who published this article.

Americas Great Resorts is a hotel marketing company, and its website promotes AI visibility services and reporting for hotels.

That doesn’t make the argument wrong. But it does mean the article is also positioning a business problem for which the publisher offers a solution.

In other words, it is both an industry observation and a marketing message.

I would want to see independent evidence behind its claims about how frequently AI recommends particular hotels, which sources influence those recommendations, and how much resulting demand converts into bookings.

My overall verdict

Strong strategic argumentI think this is one of those articles that hotel executives should read, discuss and take seriously.

But I would change its central message slightly:

The problem isn’t that Expedia bought the cafeteria.

The problem is that hotels spent years fighting over who gets the booking without paying enough attention to who controls the discovery process that happens before the booking.

That is the bigger strategic issue.

And I think there’s an even more important implication:

In the traditional distribution model, hotels worried about paying commissions to intermediaries. In the AI discovery model, they must also worry about becoming invisible, being inaccurately represented, or being excluded from the consideration set altogether.

That could ultimately be more damaging than losing a booking to an OTA.

My bottom line: I agree with the urgency of the article, but I would not respond by simply buying another AI marketing service. I’d respond by developing a measurable AI discoverability strategy that connects visibility to actual revenue and direct customer acquisition.

One question I would put to any hotel CEO or marketing director is:

If a prospective guest asks ChatGPT to recommend the five best hotels in your destination, do you know whether your property makes that list, why it does (or doesn’t), and what information influenced the answer?

If the answer is no, there’s a meaningful strategic blind spot.

And that’s the part of this article I think deserves the most attention.

Gemini Response

The article “The Bully Bought the Cafeteria” from Americas Great Resorts uses a sharp playground metaphor to critique how the hospitality industry handles online travel agencies (OTAs)—specifically giants like Expedia.

The core thesis is captured neatly in its opening premise: The hotel industry thought it was winning a war of attrition by negotiating better terms, managing commissions, or driving small percentages of direct bookings, while the dominant OTAs quietly consolidated structural control over the entire distribution ecosystem.

Here is a breakdown of the key themes in the article and an analysis of its perspective:

1. The Core Metaphor: Tactical Wins vs. Structural Defeat

• The Playground Dynamic: In the article’s framing, hoteliers spent years fighting individual skirmishes with “the bully” (OTAs)—quibbling over commission rates, parity clauses, or minor contract concessions.
• The Blind Spot: Hotels celebrated these small tactical victories (“saving lunch money” or negotiating a fraction of a point off a commission) as proof that they were managing OTA dependency. Meanwhile, the platform itself bought the cafeteria—meaning they built, bought, or locked down the underlying infrastructure where consumer intent and guest discovery actually happen.

2. The Illusion of Independence

The article targets a dangerous form of complacency in hotel marketing. Many independent hotels and luxury properties believe that as long as they run a sleek booking engine on their own website and squeeze out a few more direct bookings, they are masters of their own destiny. The piece argues that this is an illusion: if the overarching discovery layer (where travelers search, compare, and filter hotels) is entirely owned and intermediated by a handful of tech behemoths, the hotel is still fundamentally a tenant in someone else’s building.

3. The Broader Shift (OTAs to AI-Mediated Discovery)

When viewed alongside the site’s broader body of work on AI discovery and “Knowledge Formation Optimization” (KFO), the article serves as a warning about systemic shifts. Just as hotels missed the structural consolidation of OTAs decades ago, the author argues the hospitality sector risks making the exact same mistake with AI-driven search and answer engines. If AI models and aggregators dictate how a luxury resort is categorized, recommended, or summarized to a high-intent traveler, hotels lose control at the demand-origin layer, long before the guest ever reaches an OTA or a direct website.

My Perspective on the Article’s Argument

• It’s a sobering reality check: The hospitality industry has historically suffered from short-termism, focusing heavily on occupancy metrics and quarterly RevPAR (Revenue Per Available Room) while ignoring who actually owns the customer relationship and data. The “cafeteria” metaphor effectively highlights how owning the channel trumps winning individual negotiations.
• The strategic takeaway: The article is ultimately a pitch for brand autonomy and foundational demand control. It suggests that superficial digital marketing (like running standard pay-per-click ads or basic social media campaigns) cannot protect a property if its underlying market identity is entirely dependent on third-party intermediaries.

Ultimately, it’s a punchy, provocative wake-up call for hotel executives and luxury property marketers to look past daily operational metrics and look closely at who actually controls the pathways to the modern traveler.