Artificial intelligence is no longer sitting politely at the inspiration stage of travel. It is moving into hotel search, booking funnels and customer assistance—the places where recommendations become revenue and a vague idea about “somewhere warm” becomes a transaction.

Allianz Partners’ latest U.S. findings say 37% of American travelers are using or planning to use AI to help organize summer vacations. The figure comes from its 2026 Global Travel Confidence Index, conducted online by Ipsos among 11,010 adults in ten markets, including 2,001 respondents in the United States between March 20 and April 14. Travelers reported using AI for destination research, restaurant and activity ideas, itineraries and budgeting.

The interesting part is not that people ask chatbots for recommendations. They have done that for several travel seasons. What matters now is that major hotel groups are building themselves directly into conversational search, turning AI from a planning novelty into a new distribution channel.

AI is moving closer to the booking

Hilton introduced its AI Planner in March as a generative-AI concierge on Hilton.com. The beta tool lets travelers describe the trip they want, compare properties and explore amenities through conversation rather than fixed filters. Hilton initially limited access, then said the planner was available to all Hilton.com visitors from March 17.

Wyndham followed in May with a native ChatGPT app covering approximately 8,400 hotels. Users can refine results with natural-language prompts, maps, amenity filters and interactive hotel cards before moving to WyndhamHotels.com to complete the reservation.

READ MORE: Free eSIMs Now Included With CoverForYou Travel Policies

IHG launched its own ChatGPT app in June, offering real-time availability, pricing, maps and amenity information across more than 7,000 hotels in over 100 countries. It also announced conversational search for IHG.com and the IHG One Rewards app, although that part was still described as “coming soon.” Marriott entered the same race with Ask Bonvoy, an AI-powered hotel and destination search experience launched in beta in June and now visible on Marriott’s website.

Taken together, these are not four unrelated chatbot experiments. They show hotel groups trying to control how their inventory is interpreted, recommended and ultimately booked when the first search box is no longer a conventional search box.

The new fight is over the answer

Traditional hotel search begins with dates, destination and occupancy. Conversational search can begin with something far messier: “Find me a quiet hotel near the metro, with a proper desk, late check-in and somewhere decent to eat after 10 p.m.”

That is useful because real travel decisions are full of preferences that do not fit neatly into dropdown menus. It also gives hotel companies a chance to present their own property data, loyalty benefits and live rates before an intermediary defines the shortlist.

There is, however, a practical limit hiding beneath the polished demos. Most current tools still hand the traveler to the hotel group’s website to finish booking. The AI may narrow the choice, but the transactional plumbing remains largely conventional. That handoff is not a minor detail: every extra step is a place where availability can change, context can disappear or the customer can simply leave.

Accuracy will matter more than personality. A charming assistant that invents a pool, misunderstands a cancellation condition or recommends a property unavailable on the requested dates is not a concierge; it is an expensive source of complaints. Marriott’s own terms explicitly describe Ask Bonvoy responses as AI-generated outputs, a reminder that conversational ease does not remove the need to verify the result.

Travel insurance exposes the real priority

The Allianz research also offers a useful reality check. Among travelers across the ten surveyed markets who planned to travel with insurance, 63% selected 24/7 customer support and emergency assistance as an expected or valued insurer-app feature. In the U.S. sample, the figure was 53%. Digital policy management and claims tools also ranked highly.

By comparison, an eSIM function was selected by 16% globally and 21% in the United States. That does not make connectivity irrelevant. It shows that bundling an eSIM into an insurance app is a useful extra, not a substitute for the service travelers need when a flight is cancelled, luggage disappears or medical help is required.

READ MORE: AI in Hotel Distribution: How Agentic Hospitality Empowers Direct Bookings

For insurers, the sensible role of AI is therefore less glamorous than producing dreamy itineraries. It can triage requests, translate conversations, surface policy information and provide status updates. But the winning experience will still need fast escalation to a competent human when the situation stops being hypothetical.

Conclusion: AI is becoming travel’s new shelf space

The 37% figure is not proof that Americans are ready to delegate an entire holiday to an autonomous agent. It does show that conversational tools are becoming a normal starting point for travel decisions.

For hotel groups, that creates a distribution contest: be accurately represented inside AI-led discovery, preserve the route to direct booking and prove that conversational search converts better than a clever demonstration. For insurers, it means using automation to shorten the distance to real assistance, not placing a chatbot between a stranded traveler and help.

AI may write the shortlist. The brands that win will be the ones that keep the information trustworthy and make the next action work.