71 Percent of Travelers Want AI to Book Their Trips — But Only If They Stay in Control

You tell an AI assistant: "Book me a boutique hotel in Kyoto for my parents' 50th anniversary. They need a ground-floor room, walking distance to a temple, with a bathtub, under $200 a night." The AI finds three options, compares cancellation policies, checks accessibility reviews, and presents a recommendation. You review it, approve it, and the booking is confirmed. That is the experience 71 percent of American travelers say they want, according to a study published this week by Dune7 and Flesh & Bone.

The online study of 1,000 U.S. adults who had flown in the past 12 months was fielded March 6-9, 2026 and released April 8 via PRNewswire. The headline finding reframes a conversation that travel executives have been having for two years: consumers are not rejecting AI booking. They are rejecting the version of AI booking the industry is currently building.

What travelers actually want

Asked about an AI assistant that can search, compare, select, and book travel based on personal preferences, 71 percent of respondents said they were interested. The most popular tasks were booking hotels (66 percent), booking flights (65 percent), and assembling personalized travel packages (61 percent). Interest was highest among Millennials, business travelers, frequent international travelers, and people already using AI tools in daily life.

The conditions attached to that interest are not negotiable. The top concerns across the sample were AI errors being hard to reverse, unclear accountability when something goes wrong, lack of human support at critical moments, and personal data privacy. Tom Buckley, cofounder of Dune7, summarized the finding in one sentence: "The market is not saying 'don't let AI book for me.' It is saying, 'let AI do the work — but inside rules I set, with approval rights, transparency, and a human fallback when it matters.'"

That is a different product from the one most travel companies are building.

The trust gap nobody wants to talk about

Separate research from Skift earlier this month put the number of U.S. consumers willing to use fully autonomous AI booking agents, with no human oversight at all, at 2 percent. In the same report, 80 percent of travel executives said they plan to deploy AI agents at scale. Skyscanner founder Gareth Williams put it bluntly in a March interview: "I've been really struck by how negative the public is towards AI compared to people inside the industry."

A January 2026 Global Rescue survey of experienced travelers found 79 percent uncomfortable with AI making autonomous booking decisions, and only 22 percent willing to rely on AI for their next international journey.

The infrastructure side keeps moving. Google's AI Mode now executes agentic bookings for restaurants in the U.S. through OpenTable, Resy, and Tock, for event tickets via Ticketmaster, StubHub, and SeatGeek, and for beauty and wellness appointments through Booksy, Fresha, and Vagaro. Flights and hotels are next, with Booking.com, Expedia, and Marriott integrated as launch partners. IDC projects that up to 30 percent of travel bookings may be executed by AI agents by 2030.

What this means for travel platforms

Hotels and airlines whose inventory is not machine-readable will be invisible to AI agents within two years. That part of the industry message is correct. What the 71-vs-2 gap reveals is the other half: being machine-readable only solves the supply problem. The demand problem, which is consumers wanting a copilot rather than an autopilot, needs a different answer.

No clear accountability framework exists for who is liable when an AI agent books the wrong hotel, misses a cancellation window, or charges a card against stale preferences. The Dune7 study makes clear that the biggest barriers to adoption are trust-related, not technology-related. Building faster autonomy without building that trust layer first is the opposite of what the research is asking for.

Mubboo's Take

The 71 percent figure tells us something important: consumers want AI to do the tedious work of searching, comparing, and narrowing down options. The 2 percent figure tells us something equally important: almost no one wants AI to make the final decision without their approval. That gap, between wanting AI to help and trusting AI to act, is exactly where independent comparison platforms operate. At Mubboo, our Travel channel is designed for the moment after AI narrows the options but before the consumer commits. We provide the editorial judgment, the scenario-specific advice, and the honest assessment that AI agents cannot generate on their own. The 71 percent wants AI efficiency. The 98 percent wants human judgment before the credit card is charged. Both are right.