Here’s What Happens When AI Starts Making Decisions and Bookings for Hotel and Travel Guests

It’s becoming difficult to get through a day of hotel or travel industry news without hearing about agentic AI.

And for good reason. The technology has the potential to fundamentally change how guests discover destinations, choose hotels, make bookings, manage trips and interact with travel companies before, during and after their stay.

But its potential to improve the guest experience is far from guaranteed.

Guests are already turning to AI to research destinations, compare hotels, build itineraries and make sense of increasingly complex travel choices. Booking.com’s Global AI Sentiment Report found that 89% of consumers want to use AI in future travel planning, with AI assistants already considered a more trusted source of travel information than bloggers or social media influencers.

Other research points in the same direction. McKinsey found that while fewer than one-third of travelers had used generative AI for travel-related tasks, 84% of those who had said it improved their experience.

That growing comfort with AI is changing expectations.

A guest who can ask ChatGPT to recommend a quiet boutique hotel near good restaurants, create a four-day itinerary around personal interests or compare destinations based on budget and weather is unlikely to be satisfied returning to digital experiences built entirely around filters, dropdown menus and static FAQ pages.

Travel brands are responding.

Booking.com has steadily expanded its AI Trip Planner and introduced tools including Smart Filter, Property Q&A and Review Summaries to help guests search for accommodations in natural language, understand properties and make decisions more quickly. The company has also moved further into agentic AI with tools designed to streamline communication between guests and accommodation partners.

Hilton followed with the Hilton AI Planner in March of this year, a generative AI-powered conversational concierge that allows guests to describe the type of trip they want, explore destinations, compare Hilton properties and investigate amenities.

And the AI evolution is moving beyond inspiration and search. Agentic systems increasingly have the potential to help with post-booking activities ranging from reservation changes and cancellations to service requests and disruption management.

The question for the industry is no longer whether AI will become part of the guest journey.

It is how much of that journey it should control.

From Searching and Selecting to Asking and Acting

Agentic AI is moving travel from an industry built largely around searching and selecting toward one built around asking and acting.

The difference matters.

A traditional hotel booking journey might require a guest to enter a destination and dates, apply filters, compare properties, open multiple pages, read reviews and determine which option best fits the trip.

An AI-powered travel experience can start with a request as natural as:

"I’m traveling to New York with my teenage daughter for four nights. We want somewhere safe and walkable, close to shopping and Broadway, but not directly in Times Square. I’d like a hotel with two beds and breakfast for less than $500 a night."

The system can interpret the intent behind the request, consider multiple criteria simultaneously and present a shortlist.

Agentic systems take that another step.

Rather than simply recommending what the guest should do, they can increasingly take actions on the guest’s behalf: checking availability, modifying reservations, answering property questions, requesting information or coordinating multiple elements of a trip.

That changes the competitive landscape.

Hotel companies, OTAs, airlines, TMCs and other travel providers will increasingly have to decide how much of the guest relationship they want to own themselves and how much they are prepared to hand over to third-party AI platforms.

Just as importantly, guests will begin comparing every travel experience with the best AI experience they receive anywhere else.

Conversational interaction, personalization and immediate answers will gradually stop feeling innovative.

They will simply become expected.

AI Is Blurring the Line Between Service and Revenue

AI is already changing the earliest stages of the travel journey.

Guests are increasingly using generative AI for destination inspiration, hotel research, itinerary creation and recommendations. Phocuswright reported that nearly four in ten U.S. travelers used generative AI when researching trips in 2025, an increase of 11 percentage points in one year.

The bigger change will happen when those interactions become connected.

Today, travel remains highly fragmented.

A guest may research a destination through ChatGPT, search hotels through Google, compare properties on an OTA, book directly with a hotel, message the property about an airport transfer, ask the concierge for restaurant recommendations and later contact support about a billing issue.

Agentic AI creates the possibility of connecting many of those interactions.

The guest does not necessarily care which PMS, CRM, booking platform, messaging system or service application sits underneath the experience.

They simply expect the company to understand who they are, what they have already asked for and what needs to happen next.

That begins to blur traditional boundaries between marketing, reservations, operations, service and revenue generation.

A guest asking whether a resort has adjoining rooms may appear to be making a simple service inquiry. But the answer could determine whether a $6,000 family vacation gets booked.

Someone asking about spa availability before arrival may become an incremental spa reservation.

A business guest asking to extend a stay may create another room night.

A request for dinner recommendations may become a restaurant booking.

The conversation itself increasingly becomes part of commerce.

Over time, AI could make the guest relationship a continuous interaction, extending from inspiration and booking through pre-arrival, the stay itself, disruption management and post-trip support.

Guests Still Don’t Want to Give Up Control

Guests are clearly becoming more comfortable with AI, but the industry should be careful not to interpret usage as permission for unlimited automation.

There is an important distinction between guests wanting AI assistance and guests wanting AI to take complete control.

Expedia Group’s 2026 research describes what it calls an “AI Trust Gap.” Travelers may embrace AI for planning and discovery, but nearly 70% still prefer to complete bookings through trusted travel brands rather than AI chatbots or agents.

That distinction should influence how travel companies deploy agentic technology.

Guests are likely to welcome AI when it removes work: finding suitable hotels, comparing options, explaining policies, checking availability, modifying straightforward reservations or answering repetitive questions.

Their enthusiasm may decline considerably when an AI system makes a consequential decision they did not understand or authorize.

The goal should therefore not be maximum automation.

It should be appropriate automation.

The strongest implementations will understand when a guest wants convenience, when they want control and when they want another human being.

Travel Has a Much Lower Tolerance for AI Getting It Wrong

Travel is particularly unforgiving because the consequences of an AI error can be much greater than in many other industries.

A poor product recommendation is frustrating.

A nonexistent hotel, incorrect visa requirement, missed airport transfer, wrongly cancelled reservation or bad rebooking decision can disrupt an entire trip.

And many travel purchases carry considerable emotional and financial weight.

A honeymoon, family vacation, anniversary trip or important business meeting cannot simply be reordered the following day.

That makes accuracy fundamental to trust.

Guests may forgive an imperfect chatbot response. They are much less likely to forgive an AI system that confidently provides incorrect information or takes the wrong action.

And the more autonomy travel companies give AI, the higher that standard becomes.

An AI system that recommends something is useful.

An AI system that acts on behalf of the guest carries responsibility.

AI Can’t Fix a Fragmented Experience

One of the biggest mistakes travel companies can make is treating AI as another interface sitting on top of an already fragmented technology stack.

Guests do not want another chatbot that cannot see their reservation, does not recognize their loyalty status and ultimately responds with, “Please contact the hotel directly.”

AI becomes useful when it can work across the systems required to resolve the request.

A hotel AI assistant, for example, may need access to property information, reservation data, loyalty profiles, amenities, restaurant availability, transportation options, policies and messaging systems.

A travel management AI may need to understand guest preferences, corporate travel policies, budgets, approval requirements, flights, hotels, disruptions and duty-of-care considerations.

The intelligence of the model matters.

But so does the information underneath it.

That presents another challenge: travel information changes constantly.

Hotels change offerings. Amenities close temporarily. Resort fees change. Restaurants update menus. Policies evolve. Attractions close. Flights change. Guests change their preferences.

There will be no permanent "set it and forget it" version of an AI-powered guest experience.

The knowledge, integrations and performance behind it will require continual monitoring and improvement.

The Best Use of AI in Travel & Hospitality

One of the most important questions surrounding agentic AI is what happens to the human role.

In travel and hospitality, that role remains critical.

Hilton reports that 61% of travelers find AI tools valuable when planning trips. At the same time, its London concierges are increasingly being asked by guests to validate or improve itineraries originally suggested by AI.

That may offer a useful foreshadowing of where hospitality is heading.

AI does not necessarily eliminate the concierge. It may actually make the concierge more valuable.

Routine questions about breakfast hours, Wi-Fi, directions, checkout policies, airport transportation or basic restaurant options can increasingly be handled automatically.

That gives hotel employees more time to focus on the things technology still struggles to replicate: understanding nuance, recovering a difficult service situation, recognizing emotion, making judgment calls and creating memorable experiences.

The same applies elsewhere in travel.

AI can handle repetitive booking verification calls so operations teams can concentrate on exceptions.

It can answer straightforward policy questions so travel advisors can manage complicated itineraries.

It can summarize a guest’s previous interactions so a hotel employee immediately understands the context behind a complaint.

This is where some of the greatest value of agentic AI may ultimately emerge.

Not from replacing people, but from removing the repetitive administrative work that prevents talented people from spending more time on the experiences where they add the most value.

The handoff between AI and people will therefore matter enormously.

When a guest needs a person, that employee should not have to begin with:

"Can you explain what happened?"

They should already know. 

BTW - This is a major frustration of mine when dealing with any organization. I hate having to repeat everything multiple times,only to get no resolution at the end!

How Far Will Agentic Travel Go?

Travel already requires people to surrender a lot of control.

They hand companies their money, identity information, credit cards, schedules and sometimes thousands of dollars months before receiving the actual experience.

Agentic AI asks them to surrender something else:

Decision-making authority.

That makes trust more important than ever.

The travel and hospitality industry needs technology which consistently demonstrates that it understands the guest/traveler, works from accurate information, knows its limitations and brings a person into the experience when necessary.

Speed, convenience, personalization will matter.

But travelers need to feel that the technology is working for them rather than simply reducing work for the company deploying it.

That may ultimately be the dividing line between AI that strengthens the guest experience and AI that damages it.

Guests will tolerate technology and increasingly embrace automation.

What they will not tolerate for long is technology they cannot trust.