A family of four in Atlanta asking an AI chatbot for a spring break idea and a solo traveler in Chicago working with a tight budget might expect two very different answers. Yet, new findings from Skift Research suggest they would mostly get two different sales pitches for a similar set of places.
For its report, “The Mystery of the Disappearing Destination: How AI Is Transforming Travel Discovery,” Skift Research ran repeated prompts through ChatGPT, Gemini, and Claude across traveler personas that varied by demographics, interests, travel companions, and budget. The chatbots adjusted their language to each person, offering family-friendly activities to parents and low-cost flights to travelers with less money, while the destinations themselves barely shifted.
Skift summed up its findings in the article’s editorial takeaway: “For now, AI personalizes the message, but it doesn’t personalize the recommendation.” The five destinations that came up most often accounted for 38% of all AI recommendations, close to the roughly 40% share those same five hold of actual global inbound arrivals. In practice, a traveler who opens a chatbot hoping for something new is often handed a short list dominated by places that already draw the biggest crowds on earth.
Why AI Travel Recommendations Keep Landing On The Same Places
When Skift Research and McKinsey surveyed U.S. travelers about how they like to shop for trips, 63% of the 1,015 respondents chose a curated shortlist of 10 options over a 500-option catalog (30%) or a single editable AI-generated itinerary (7%). That preference shows travelers want the narrowing chatbots provide, even as that narrowing leaves most of the map off the page. A separate Skift report on Chinese-language AI travel answers helps explain how the shortlist forms.
Marketing firm Create Consulting China tested 240 travel queries across five Chinese AI platforms, including DeepSeek, Baidu’s ERNIE Bot, and ByteDance’s Doubao, using eight traveler profiles. Open-ended questions clustered around Singapore, Thailand, and Japan, and the study found AI favored destinations with easy visa access, short flight times, strong safety records, and plenty of Chinese-language content, producing a feedback loop that strengthens places already winning attention. Destinations with more written about them get recommended more, which brings more visitors, who then produce more content for the next model to draw on.
How To Find Destinations AI Travel Tools Overlook
In that same study, question wording shifted the results considerably. When researchers limited queries to a specific continent, South Africa and Mexico appeared in the answers, even though neither showed up in any of the open-ended results. For U.S. travelers, the same approach means asking a chatbot about a specific region, month, or flight length from a home airport instead of asking where to go in general.
Fare-based tools approach discovery differently. Google Flights Explore plots prices from a traveler’s departure airport across a world map, and Thrifty Traveler notes that users can cap the map to show only results under a set price. The same guide flags the tool’s quirks, noting that results depend on how the map is moved and are not always sorted by price or highlighted on the map, so dragging and zooming across regions surfaces airports a default view leaves out.
Local Publications, Tourism Boards And The Towns Beyond AI’s Shortlist
Skift Research’s advice to destination marketers points travelers toward the sources that feed AI in the first place. The report urged destination marketing organizations to treat AI visibility as a distribution problem, improve structured local content, and work with the operators and platforms that supply information to AI systems.
Official tourism board websites and the local operators they list sit at the start of that chain, and a traveler who reads them directly sees festival calendars, regional routes, and small businesses before any model condenses them into three suggestions. City magazines, alt-weeklies, and regional newspapers fill in the same picture on a weekly cycle, covering openings and neighborhood events that may take time to build the volume of online content AI tends to favor.
Stiplo’s research on AI visibility focuses on hotels instead of towns, drawing on 27,360 AI travel answers about 659 properties in Barcelona, which the company presented at Skift’s Data + AI Summit in London. Independent hotels were eight times more likely than chains to be unknown to an AI system, yet once a system knew a property existed, it recommended independents at about the same rate as chains. Skift’s coverage framed the gap as a discovery problem instead of favoritism toward chains.
