Travel companies now use AI in four main places: planning assistants that suggest trips, pricing and revenue tools, customer-service agents, and operations such as flight routing. This article stays close to what the companies themselves have published, including Expedia, Booking.com, Airbnb, Delta and Alaska Airlines, and also covers where the technology falls short. For a broader overview, see our explainer on how AI is changing travel and hospitality; this piece focuses on named, documented examples.
Trip planning assistants
The first wave of conversational planners arrived in 2023. On April 4, 2023, Expedia announced a beta that let people chat in its iOS app, in English, about destinations, lodging, getting around and activities. Hotels mentioned in the conversation were saved to a trip so users could check dates and availability and add flights, cars or activities. Expedia’s own release was frank about the limits: it called the product a beta, said results could be inaccurate or responses inappropriate despite safeguards, and said it would review feedback to improve answers.
The pattern has since moved into general-purpose chat assistants. On May 20, 2026, Booking.com described an app inside ChatGPT for car rentals: a traveler describes what they want, such as an automatic SUV in Mallorca, browses the options returned and is sent to Booking.com to finish the reservation. In both cases the assistant handles the discovery, while the transaction stays on the travel company’s own platform.
Do travelers want more than that? Expedia Group commissioned a YouGov survey of more than 5,700 adults in the US, UK and India, run March 10 to 25, 2026. It found 53% comfortable with AI suggesting travel options and 40% using AI to help build itineraries. Only 8% of US and UK travelers said they rely on chatbots or agents when planning, versus 59% using search engines. When it came to paying, 68% preferred booking with a trusted travel brand, and 66% said they would not trust an AI assistant to buy or book anything for them. This is company-commissioned research, so treat the exact figures as indicative.
Pricing and revenue management
Dynamic pricing is older than the current AI wave, but newer tools make it more visible and more contested. Delta has told US lawmakers that it is evaluating an AI pricing recommendation tool developed by Fetcherr in domestic and international test markets. According to Delta’s letters, the tool is meant to reduce manual work for revenue-management analysts, who review and adjust its suggestions, and it recommends both increases and decreases.
The more important point for travelers is what data goes in. Delta says the tool receives aggregated route-level demand and market information and no customer personal data, and that its filed fares “do not take any customer’s personal data into account.” Those are Delta’s statements, not independent findings. They show that the debate has moved from whether airlines use algorithms to whether prices are targeted at individuals, which is why price transparency is now a regulatory topic as well as a consumer one.
Customer-service agents
Support is where companies report the clearest numbers. Airbnb’s second-quarter 2025 shareholder letter said it was expanding an AI-powered customer service agent to 100% of US users and that the share of hosts and guests who needed to contact a human agent fell by 15%. The letter described the rollout as a step in perfecting its core service. Note what the figure measures: fewer human contacts, not a verified improvement in guest satisfaction.
Booking.com’s May 2026 release adds smaller examples. It lists an AI Car Rental Helper that answers questions about payment methods, pickup checklists and insurance around the clock, and AI-generated summaries of recent traveler reviews for car rental locations. The company said early testing reportedly led to fewer support calls and disrupted pickups, again a company-reported result.
Operations and disruption handling
Some of the most practical uses never face a customer. In 2021 Alaska Airlines announced a multi-year contract for Flyways, an AI platform from Airspace Intelligence that gives dispatchers a map of air traffic and recommends reroutes around weather and congestion. In the six-month trial, the airline said Flyways found a chance to cut miles and fuel for 64% of mainline flights and that dispatchers accepted 32% of recommendations, and the company reported savings of 480,000 gallons of fuel. Those figures come from the announcement, and the trial took place during sharply reduced pandemic-era flying. Dispatchers made the final decision, which is the common design in aviation.
This is routing, not passenger rebooking, and we did not find an official airline page that documents AI-driven rebooking numbers. Readers should be skeptical of any claim that goes further than a company’s own published figures.
Limits and risks
- Wrong answers carry legal weight. In Moffatt v. Air Canada, decided on February 14, 2024, a British Columbia tribunal found the airline responsible for incorrect refund guidance given by its website chatbot and awarded the customer CA$812 plus costs. The tribunal rejected the idea that the chatbot was a separate entity. See our glossary entry on hallucination for why such errors happen.
- Itineraries can be invented. Language models can state opening hours, visa rules or distances with confidence and be wrong. Expedia’s own beta notice said results can be inaccurate.
- Pricing can be opaque. Even when a company denies personalized pricing, travelers usually cannot see how a fare was set. Comparing dates and sources remains the best defense.
- Human oversight is the norm. Delta says analysts review pricing suggestions, and Alaska dispatchers decide on reroutes.
If you plan trips with these tools yourself, our step-by-step guide to planning a trip with AI lists checks to run before you book.
What to watch next
- Whether assistants inside chat apps begin completing bookings and payments instead of handing off to travel sites.
- Regulatory attention to airline and hotel pricing algorithms, including what data they may use.
- More published results from service agents, ideally with satisfaction measures and not only contact reduction.
- Disclosure standards for AI-written itineraries and review summaries.
Key takeaways
- Companies use AI for planning, pricing, service and operations, but most published results are self-reported.
- Chat assistants currently discover and recommend; booking usually still happens on the travel brand’s platform.
- Delta says its pricing tool uses aggregated route data and that analysts review its recommendations.
- Airbnb reported a 15% drop in the share of users needing a human agent after its AI rollout.
- Companies remain accountable for what their chatbots say, as the Air Canada ruling shows.
Frequently asked questions
How do travel companies use AI today?
They use it for conversational trip planning, pricing recommendations, customer-service agents and operations such as flight rerouting. Examples include Expedia’s in-app planner, Booking.com’s car rental helper, Airbnb’s service agent and Alaska Airlines’ Flyways platform.
Is AI deciding the price of my airline ticket?
Delta says its AI tool only recommends fares to human analysts and uses aggregated route-level data rather than personal information. Other airlines may differ, and these statements are company claims.
Can I trust an AI itinerary?
Use it for ideas and structure, but confirm opening hours, visa rules and prices on official sites. Expedia itself warned that results from its AI beta can be inaccurate.
