AI travel recommendations personalized to a traveler’s budget, age or travel companions often point to the same handful of destinations no matter who is asking, according to research from Skift Research. The firm tested ChatGPT, Gemini and Claude with repeated prompts built around different traveler personas, varying demographics, interests, companions and budgets. The pitch changed. The places did not move much at all.
That gap matters because more travelers are leaning on chatbots to shortcut the planning process, trusting that a tailored-sounding answer means a tailored result. The data suggests otherwise, and a parallel trend shows a growing number of travelers pairing AI tools with human advisors or ditching algorithms altogether for analog trip planning.
The Numbers Behind AI Travel Recommendations Personalized By Persona
Skift Research found that the top five destinations suggested by the AI models accounted for 38% of all recommendations, close to the 40% share those same destinations held in actual global inbound arrivals in 2024. The top ten AI-recommended destinations made up 48% of suggestions, compared with 56% of real-world arrivals, according to the research.
Those figures show the AI models were not inventing obscure destinations tailored to niche travelers. They were largely reflecting where people already go, dressed up in persona-specific language.
What Actually Changes: Activities, Not Places
The research found AI models personalize the surrounding pitch and suggested activities far more than the destinations themselves. Families prompted the chatbots and received family-friendly activity suggestions. Budget-conscious travelers got pointed toward low-cost flight options. But the underlying destination recommendations barely shifted between personas, according to Skift Research.
In practice, that means two travelers with very different profiles asking the same AI tool “where should I go” are likely to land on similar shortlists, even if the supporting details around each pitch read as custom-built.
Travelers Still Want A Human Check
The appetite for a curated, limited set of options rather than an overwhelming catalog came through clearly in consumer preference data. Skift Research found 63% of 1,015 U.S. travelers surveyed preferred a curated shortlist of 10 options over either a 500-option catalog or a single AI-generated itinerary.
Separate research from ALG Vacations reinforces that travelers are not ready to hand full control to an algorithm. Among people who use AI to plan a trip, 77% still want a human to make the final call, according to ALG Vacations. The same research found 51% of millennials and Gen Z travelers who use AI have also booked through a travel advisor, pointing to a hybrid approach rather than a wholesale replacement of human expertise.
Travel advisors describe their value in exactly the areas where generalized AI output tends to fall short. Jacquie Whitt of Adios Adventure Travel said travelers often handle straightforward bookings online but still seek out advisors for the harder calls. “Although travelers seem to like the convenience that digital offers for making online reservations, they contact me for advice to help organize their tours, activities and to get information about special documents,” Whitt said. She cited guidance on choices such as whether a 500-guest or 140-guest ship suits an Antarctica expedition better, a judgment call shaped by experience rather than aggregate data.
The Counterpoint: AI Tools Are Getting More Useful For Advisors
The strongest case against treating generalized AI output as a dead end comes from the advisor side of the industry, where AI is being built directly into booking workflows rather than used as a standalone planning tool. TL365 developed an AI solution that lets travel advisors search for hotels using natural-language prompts inside a chat environment, according to Travel Weekly. Ian Cambata, CEO of TL365, explained the reasoning behind keeping search and booking in one place. “If the advisor has to spend four days of research outside of the platform to find the right room to then come back and book, workflow-wise, it doesn’t make a lot of sense,” Cambata said.
TL365’s system also includes a feature that keeps customer data and personally identifiable information from interacting with the AI, addressing a privacy concern that consumer-facing chatbots do not typically resolve. Separately, 1000 Mile Travel Group launched the latest version of its Advisor Hub with AI-powered search, now available to advisors in the U.S., U.K. and Australia, according to Travel Weekly. In these cases, AI is narrowing research time for professionals who still apply their own judgment to the final recommendation, a different use case than a consumer typing a vacation idea into a chatbot and expecting a finished plan.
A Parallel Shift Toward Analog Planning
Some travelers are responding to automation fatigue by stepping away from digital tools entirely. Bob Gourley, who works for a computer security company in Washington, D.C., hired a travel advisor instead of booking online for a trip to Muscat, Oman, and brought a 35 mm camera instead of his phone. “I had 36 shots.
I stopped more. I observed more,” Gourley said, describing how the limited frame count changed his pace, according to USA Today. Mitch Goldstone, CEO of ScanMyPhotos.com, framed the broader pattern this way: “In a world that’s increasingly automated, people are craving that human connection more than ever.”
Richard Campbell, founder of 10Adventures, echoed that sentiment around discovery itself, saying, “I want to be surprised and amazed seeing something for the first time. I think this is a key part of travel.”
That appetite for unmediated experience has produced products built around it. Hotel Rangá in southern Iceland sells an Analog Menu package for 20,100 kroner, about $165, that includes two disposable cameras, a deck of playing cards, a coloring book on Icelandic history with crayons, three postcards and a hand-marked map pointing to staff favorite stops. The Moana Surfrider in Waikiki runs a One Year Wave experience in which guests write postcards to themselves that the hotel mails a year later.
Architect Laura Intscher sketches ancient ruins rather than photographing them, saying she notices more detail and remembers places more vividly as a result, according to USA Today. The USA Today reporting also noted an increase in travelers younger than 40 turning to advisors for trip advice.
What To Watch Next
The Skift Research findings suggest travelers who want a genuinely tailored destination list, not just tailored activity suggestions, should treat AI output as a starting point rather than a finished answer. Cross-checking a chatbot’s top picks against independent sources, or running the same prompt with different personas to see how much the destination list actually shifts, is one way to spot generalized recommendations dressed up as custom ones. The hybrid pattern identified by ALG Vacations, where AI users still book through advisors at meaningful rates, points to where the industry is heading next: AI tools embedded in advisor workflows for research speed, paired with human judgment for the decisions that carry the most weight. Further platform rollouts from companies such as TL365 and 1000 Mile Travel Group will show whether that combination becomes standard practice across the advisor market.
Don't stop here
Every Score. Every Story. One Home.
Breaking sports, entertainment and news from across the Total Apex network, updated all day.
Get the Full Story →
