--- title: "How to Get Your Hotel Recommended by AI: 36% of Korean Travelers Already Ask ChatGPT" description: "TL;DR In May 2026, Google, OpenAI, Anthropic, Booking.com, Expedia, Airbnb, and TikTok all pointed the same direction. Hotel bookings are moving inside AI. When guests know your hotel's name, AI recommends it 97% of the time. When they don't and ask for suggestions, your odds drop to 10–15%. Your pipeline of new guests is closing. 36% of Korean travelers already use AI to plan trips—faster adoption than the US (25%). The way forward is clear: make AI..." published: 2026-05-26T09:00:00+00:00 author: "ONDA 편집팀" category: "Insights" image: "https://zqcfqfqgiyckyhazcfrk.supabase.co/storage/v1/object/public/blog-images/webflow/6316f3f75130788ec2d762dd/6a16a8f012e768c86f1d657a_2026-05-26-img1-_.png" canonical: https://onda.me/en/blog?slug=how-to-get-your-hotel-recommended-by-ai locale: en --- # How to Get Your Hotel Recommended by AI: 36% of Korean Travelers Already Ask ChatGPT ## May 19, 2026 — Google I/O **TL;DR** - In May 2026, Google, OpenAI, Anthropic, Booking.com, Expedia, and Airbnb all pointed the same direction as if on cue. **Hotel bookings are moving inside AI.** - When guests know your property's name, AI mentions it almost every time. But when they don't know you and ask for a recommendation, your odds of being mentioned drop to 10–15%. **Your pipeline of new guests is closing.** - 36% of Korean travelers already use AI to plan trips—faster adoption than the US (25%). - The takeaway is simple: **make AI recommend your property.** This piece shows you how. May 19, 2026. On the Google I/O stage in Mountain View, Vidhya Srinivasan, VP of Ads and Commerce, dropped one line that changed everything: > > "We're expanding the ability for AI to complete purchases directly within search. Hotels are first." ([Google I/O 2026](https://blog.google/intl/ko-kr/company-news/technology/search-io-2026/)) > The company that owned search just declared it was coming for bookings, too. That same day, Google unveiled an AI that works in the cloud 24/7—even while you sleep. It means guests are now outsourcing hotel search, comparison, and booking to machines that never rest. AI Mode passed 1 billion monthly users in its first year. Surprising? Not really. It was more like the inevitable finally getting a stamp. Within that same May, three big tech companies, four global OTAs, and one video platform all said versions of the same thing. **Hotel bookings are moving inside AI.** This isn't one company's bet. It's something close to industry consensus. So what does that mean for your property? And what should you do about it? That's what this piece is about. Let's start with how guests are changing. ## How Guests Find Hotels Is Changing The clearest snapshot of what this shift means for us comes from hotel marketing analyst Gil Chan's May 2026 analysis. When a guest knows your property's **name** and asks ChatGPT ("How's Shilla Hotel Jeju?"), the AI mentions it in **97% of cases.** But when the guest **doesn't know you** and asks "Recommend a place to stay with my family in Jeju," the odds a specific property gets mentioned plummet to **10–15%.** Who fills the rest? Airbnb, Booking.com, Expedia. All OTAs. The key takeaway is clear: **guests who already know you are fine. The risk is with new guests you haven't met yet.** The entry point for new guests is narrowing in the AI era. And guests are moving fast. According to a Cloudbeds report analyzing 90 million bookings, the share of US travelers using traditional search to plan trips dropped from 51% to 36% in one year. Generative AI filled that gap at more than double the rate. This shift is different from past search disruptions. Search was about matching keywords. AI reads context and intent, then **decides for the guest.** Which means the same shift can be a crisis for some properties and an opportunity for others. Results are already showing up for early movers. One hotel group that reorganized its data to be AI-readable saw direct booking revenue jump 2.1x in two months, with 17% of new bookings coming from AI channels (Hospitality Net). For some, this is a crisis. For others, it's already revenue in hand. ## "But Korea Will Be Different, Right?" According to Agoda, **36% of Korean travelers already use AI to plan trips.** That's 11 percentage points higher than the US (25%)—the birthplace of AI. The share of Koreans using ChatGPT for information also climbed from around 40% to 55% in 2025 alone (Open Survey). Another signal this isn't a fad: when ChatGPT's answer doesn't satisfy, 77% say they'll "ask ChatGPT again," while only 32% would "go back to search." "But we have Naver, don't we?" Fair point. Naver is one of the only companies in the world that defended 60%+ domestic market share against Google's global search dominance. It did so by understanding Korean language and context better than anyone. Will that strategy hold in the AI era? I want to cheer for a Korean company, but I'm cautious for two reasons. **First, the language barrier has collapsed.** Global LLMs already train on data across all languages. On April 7, 2026, X (formerly Twitter) rolled out [Grok-powered automatic translation](https://techcrunch.com/2026/04/08/x-is-rolling-out-automatic-translation-and-photo-editing-powered-by-grok/) worldwide. Korean posts now reach global users in real time. The era when Korean language gave Korean platforms an edge is ending. **Second, Naver's own LLM hasn't proven itself with Korean consumers—yet.** HyperCLOVA X, which trained on 6,500x more Korean data than ChatGPT. Its consumer service ['CLOVA X' quietly shut down on April 9, 2026](https://www.sentv.co.kr/article/view/sentv202603050088)—two days after X launched auto-translation. Naver has pivoted to B2B sovereign AI for non-English markets like Saudi Arabia. [Korea Herald](https://www.koreaherald.com/article/10665662) ran a headline around the same time: "ChatGPT narrows Naver's grip on Korean search." Naver held its ground against global players in search. Whether it can do the same in AI search is an open question. But whatever the answer, **it's risky for hoteliers to stake all their visibility on Naver alone.** Your potential guests are already asking ChatGPT and Gemini. ## An Early Adopter's View on the Speed of Change I studied computer science, started as a developer, and have spent my entire career in tech. I built a booking platform for Korean homestays abroad in 2008, then founded ONDA in 2016. For 10 years, I've worked on a single mission: using technology to help the hospitality industry. One thing I've learned over that time: **change always arrives later than expected.** The internet, smartphones—I'd feel the shift first, but it would take 1–3 years to spread around me. Not anymore. Now, when I sense a change, the world around me shifts in a matter of months. When my 70-year-old mother said "I asked AI," I couldn't believe my ears. It's not that I've slowed down. **The world has sped up.** ## Big Tech's Three Paths Somewhat surprisingly, it's not all of big tech shaking up hotel bookings—just three companies that make AI. Google, OpenAI, and Anthropic (the company behind Claude). What's interesting is that all three are targeting different positions. **Google — from search to checkout, all in one place.** Google, the original owner of "search," now wants booking and payment, too. Soon, a guest will say "family pension in Gangwon-do this weekend" to Google AI, then "book this one," and Google will handle payment. Search, booking, and payment merge into one. **OpenAI — walking directly into hotels.** OpenAI, which built ChatGPT, isn't stopping at selling tools. It's sending its own engineers inside hotels and airlines to rewire operations around AI. Once you're locked in, five years down the line your pricing, customer service—everything—sits on top of OpenAI. Hospitality Today put it this way: choosing AI isn't picking software anymore. It's choosing **an infrastructure partner** you'll live with for years. **Anthropic — quietly laying the standard.** Anthropic, which built Claude, isn't grabbing hotels directly. It's building the "connection rules" that link AI to hotels. Global chain Wyndham already connected its data this way—and now surfaces across Google, ChatGPT, and Claude *simultaneously*. It's a way to avoid lock-in. Skift calls this trend 'The Claude Effect,' noting that as AI automates bookings, the commission-taking middleman (OTAs) loses share, and more value flows to suppliers like hotels. **Whoever sets the rules first captures the game.** The three paths are different, but they all look at the same destination: **hotel bookings are moving inside AI.** ## OTA's Four Paths OTAs, by contrast, aren't moving in one direction. This matters most directly to Korean hoteliers—because a significant chunk of our revenue comes from them. **Booking.com — owning the guest touchpoint.** Booking.com scaled the share of bookings where it directly handles payment, refunds, and customer support from 22% to 70% in just a few years. It's no longer a simple intermediary. It owns the guest relationship. When AI books through Booking, the guest and the data stay with Booking. Skift noted: _"Booking is ahead of its own strategy."_ **Expedia — building a department for 'AI guests.'** Expedia opened its inventory to partners' AI for direct access, then went further: it created a marketing department aimed not at human guests, but **at AI agents**. Their insight cuts to the core: _"Humans take shortcuts—brand familiarity, loyalty. AI doesn't need shortcuts. It evaluates every option one by one."_ Competition is shifting from "name recognition" to "what makes this property different." That's something brand-obsessed hoteliers can't afford to ignore. **Asian OTAs (Trip.com, Klook, Traveloka) — cautious but fast.** They're still careful about handing booking itself over to AI. Trust is the issue. But they're moving fast on operations, customer service, translation, and recommendations. Domestic Korean OTAs likely won't be much different. **Airbnb — going the opposite way.** Airbnb took the reverse path. CEO Brian Chesky said flatly: _"AI chatbots are not the new Google."_ Rather than feed data to big tech and rely on inbound traffic, Airbnb will **pull guests directly into its own app**. Bring AI inside, but don't release yourself outside to AI. Next to those four paths, a new current is rising. According to [AltexSoft's May analysis of 'Intermediaries 3.0,'](https://www.altexsoft.com/blog/agentic-ai-vs-ota-hospitality/) AI-native aggregators like DirectBooker are emerging. They aggregate hotels' direct booking inventory in one place. The key difference from traditional OTAs: **the hotel stays the 'merchant of record'** (owner of the booking and payment). Guests remain the hotel's guests. Data stays with the hotel. A new path is forming—one that reaches AI directly, bypassing OTAs. Right now, only **2–8% of travelers give AI agents autonomous purchasing authority.** **Trust hasn't built up enough** for people to let AI buy on their behalf. But that number will grow. And before that moment arrives, the urgent task isn't the booking step—it's the **discovery step.** Making sure AI surfaces your property in its answers by organizing your information in a way it can read. ## For Some, Opportunity. For Others, Threat. The question that matters to us: **how do we survive this shift—and seize the opportunity?** ## Properties That Will Seize the Opportunity **Properties with a clear concept.** AI doesn't recommend vague properties. "Quiet ocean-view pension," "hanok that allows pets," "boutique hotel with wine cellar," "guesthouse where sunrise is visible from every room." Properties that can state their identity in one line get into AI answers. "We're good at everything" translates to "nothing special" in AI's eyes. **Properties with structured digital information.** Precise location, amenity list, pricing policy, room types, English descriptions. This information needs to live in one place, consistently. AI only cites what it can cite. Scattered information doesn't get cited. Gil Chan summed it up in his [May 24 analysis](https://www.hotelnewsresource.com/article141341.html): **"Getting visibility requires the same fundamentals that have always driven search visibility: authoritative content, structured data, external citations."** **Properties with detailed reviews.** AI cites "the abalone porridge was memorable, and the owner said it was seafood he caught himself" far more than "breakfast was good." That level of detail becomes the evidence AI pulls from. Properties that design experiences where guests use specific, vivid language—and let that language flow into reviews, social media, and external content—will have the advantage. And **all of this information needs to be online where AI can easily access it.** Handwritten guestbook reviews in your lobby or content on a website not linked anywhere? Useless. ## Properties That Will Feel the Threat **Properties with scattered, unstructured information.** Fragmented reviews on Naver blogs, photos only on Instagram, no official website, and different info on every OTA channel. AI doesn't know what to trust. And when the information conflicts, AI treats it as unreliable and won't cite it. **Mid-tier chains that relied only on brand recognition.** As Skift made clear: **"AI agents don't need shortcuts."** (No reliance on brand familiarity.) Mid-tier brands that coasted on name recognition will get shaken the most. Top luxury will survive on strong identity. Small independent properties can arm themselves with a clear concept and actually gain ground. The most vulnerable position? The middle. **Properties heavily dependent on one or two OTAs.** As outlined earlier, global AI companies and OTAs are pursuing different strategies to reshape the lodging market—and no one can say yet who'll win. Over-relying on one or two channels is like betting your entire net worth on a single stock. Diversifying channels is baseline. More than ever, running a well-maintained **official property website**—where you control your own information—matters. **Properties whose online information isn't managed.** One way LLMs judge information trustworthiness when learning from the web is **accuracy.** But if your property's info differs across OTAs and channels? AI will flag it as inaccurate and skip citing it. Use a tool like a channel manager that syncs information from one data source to multiple channels (OSMU—One Source Multi Use). ## Five Ways to Make AI Recommend Your Property The answer is simpler than you think. You just need to move fast. **1. Define your property's identity in one sentence.** The one line AI will cite? You have to write it first. Something like: "Secluded ocean-view pension, pet-friendly, breakfast included." Vague = invisible. Write down your property's strengths scattered in your head, pick the most distinctive one, and embed that same sentence across every channel. **2. Structure your information in one place.** Your official website, a single PMS, or a channel manager that syncs content (not just availability) needs to hold accurate, consistent info. If every OTA has different details, AI can't decide what to trust. Location, room count, amenities, pricing, check-in/out times, pet/smoking/parking policies—every source should say the same thing. That's when AI trusts it. **3. Design experiences that elicit specific language from guests.** Not "it was delicious," but "the abalone porridge was great." Let that phrasing flow into reviews, social posts, external content. Generic reviews can backfire. Intentionally create one memorable experience for guests, and make sure that experience gets written into the digital world. **4. Get mentioned externally.** Company websites AI already trusts, local media, travel bloggers, news. AI reads external citations as trust signals. You saying your property is great means far less than another outlet mentioning it. Pay a little attention to blogger outreach, YouTubers, media interviews you used to ignore. Small efforts stack. Their mentions can guide AI to your property. **5. Grow your own channels (omnichannel).** Keep OTA revenue flowing, but build your direct booking path and owned content channels alongside. AI will assign trust points to the content and website you've built consistently over years. Don't have one yet? It's not too late. The world now cares less about websites that look good to humans and more about websites that look good to AI. Use proven solutions to build an AI-friendly website. You need your own digital asset. These five boil down to one thing: **Showing your property well—not to people, but to AI.** It sounds grand, but it's not fundamentally different from what good hoteliers have always done. Create differentiation, organize information accurately, design guest experiences, earn external praise. The only difference now is the audience has shifted from people to AI. You just need to learn the new grammar. In this process, the role of tech infrastructure—channel managers, AI visibility tools, booking engines—will only grow. ONDA, the company I founded, helps hoteliers with digital exposure and channel integration right in the middle of this shift. It's the same direction we've been looking for 18 years. ## In the Middle of an Industry Shift May wasn't just a month of news overload. **In 18 years of watching this industry, it was the fastest transformation I've ever felt.** Remarkably, three big tech companies, four OTAs, and one video platform all said the same thing in the same month. The most dangerous stance right now is: **"Let's wait and see before deciding."** When the internet first appeared, when smartphones and mobile took over, when global OTAs entered—someone rewrote the rules during each shift and became the winner. And those victories lasted a long time. This time, it's even faster. In two months, the world might shove your property far from the opportunity. To avoid being left behind and seize the opportunity, it comes down to two things: **Surface your property well on the new paths built by big tech and OTAs, or build your own channel to meet guests directly—separate from them.** It's not hard. This time, let's catch both rabbits. Guests are already asking ChatGPT. Does ChatGPT recommend your property? If not, why not? That's where you start. --- ⚠️ All rights reserved. Cite this piece as 'ONDA (온다)' when quoting.