Quick summary
AI-powered trip planning has moved from novelty to mainstream infrastructure. Booking.com, Expedia Group, and Kayak have all deployed generative AI assistants that convert a natural-language request into a fully bookable itinerary spanning flights, hotels, and activities. Industry data from 2025 shows 68% of travelers used some form of AI tool during their booking journey — up from 23% three years earlier. The shift is accelerating, and it changes not just how you plan trips but how airlines and OTAs price them.
The upside is real: faster planning, automatic disruption handling, and fewer separate searches. The catch is that the same AI optimizing your itinerary is also optimizing revenue against you.
Travel planning is getting faster, more conversational, and considerably more opaque. Booking Holdings — parent of Booking.com, Priceline, and Kayak — has been building its Connected Trip strategy for several years, using AI and machine learning to stitch flights, accommodation, car rental, and attractions into a single itinerary managed from one interface. Expedia Group has done the same with its own AI travel assistant, live across its brands. Both are in active production, not pilot phases.
The commercial logic is straightforward: keep the traveler inside one ecosystem from inspiration to checkout. For frequent flyers, this means fewer tabs, fewer logins, and — in theory — fewer missed connections between a flight change and a hotel that no longer makes sense.
In practice, the picture is more complicated.
Booking.com’s AI Trip Planner, launched broadly in 2023 and expanded through 2024–2025, lets users describe a trip in plain language and receive a bookable itinerary drawn from the platform’s accommodation, flight, and city data. Expedia Group introduced comparable functionality across its brands. Kayak integrated AI search. In Australia, Webjet went further — launching a ChatGPT app that surfaces live flight and hotel pricing directly inside a conversation, with no waitlist and no pilot phase.
The tools are here. The question is whether they work for you or against you.
What connected trip AI actually does at the booking stage
The Connected Trip concept is more than a marketing phrase. Booking Holdings’ shareholder filings confirm the strategy links multiple travel components under one AI-managed itinerary, with cross-selling across brands as a core commercial objective. When your flight changes, the system can theoretically flag your hotel, ground transport, and restaurant reservation simultaneously — something no human agent could do at scale.
United and Lufthansa use similar AI engines to bundle ancillaries and post-booking offers, targeting seat upgrades, lounge access, and travel insurance at moments when you’re statistically most likely to buy. The timing is not accidental. These are revenue-management systems running continuously, not customer-service features.
| Platform | AI tool | Deployment status | Key capability |
|---|---|---|---|
| Booking.com | AI Trip Planner | Live (launched 2023, expanded 2024–25) | Conversational itinerary from Booking’s full inventory |
| Expedia Group | AI Travel Assistant | Live across Expedia brands | Natural-language trip planning and bookable suggestions |
| Kayak | AI search integration | Live | Conversational flight and hotel search |
| Webjet (AU) | ChatGPT app | Live (no waitlist) | Live pricing inside ChatGPT conversation window |
| United Airlines | AI ancillary engine | Active production | Bundled seat, lounge, and insurance offers post-booking |
| Lufthansa Group | AI offer engine | Active production | Dynamic ancillary bundling across group brands |
Flight deals
most people never see
Our AI monitors 150+ airlines for pricing anomalies that traditional search engines miss. Air Traveler Club members save $650 per trip per person on average: see how it works.
Each deal saves 40–80% vs. regular fares:
Why the AI showing you deals is also pricing against you
Behind every conversational interface is a revenue-management engine that has been running long before generative AI arrived. Airlines and OTAs feed these systems booking-curve data, competitor fares, and your individual behavior — searches, clicks, past trips — to decide which fare family, seat, or add-on to surface and at what price. The commercial incentive is to maximize revenue per trip, not to find the cheapest option for you. Personalization and profit optimization are the same mechanism.
Traveler reports from 2025–2026 show the gap between marketing and reality is still wide. AI chat planners handle neighborhood suggestions and rough itineraries well. They struggle with closed attractions, unrealistic drive times, and multi-airport routings that look clean on screen but are miserable in practice. Rebooking disrupted multi-carrier trips still frequently requires a human agent, despite automated messages suggesting otherwise. And the real total price — seat fees, resort charges, baggage — often only appears on the final checkout screen, buried under a conversational interface that felt frictionless until it wasn’t. Understanding how airline pricing and sales mechanics actually work makes it much harder for any AI tool to obscure the full cost.
Two signals worth watching as this develops. The EU AI Act enforcement deadlines in 2026–2027 will clarify how high-risk and profiling rules apply to consumer travel apps — if guidance lands as expected, European travelers will see new consent screens and clearer explanations of AI recommendations. If enforcement stalls, OTAs and airlines will keep pushing behavioral data use until challenged. Separately, watch Booking Holdings and Expedia Group earnings calls over the next two to three quarters: rising connected trip penetration and ancillary revenue tied to AI tools will signal even more automation of customer service, with human support becoming harder to reach on low-fare tickets.
How to use AI trip tools without getting played by them
AI planning tools are genuinely useful — but they’re built to serve the platform’s revenue goals alongside yours, and the mistakes travelers make with them are consistent and avoidable.
- Run a parallel check before paying. Use an AI planner to build your itinerary, then manually re-create the same routing on at least one airline site and one metasearch tool. Compare total price including bags, connection quality, and change rules. AI-suggested layovers sometimes fall below minimum connection times for the airport in question — check those independently before committing.
- Expand fare details before you close the chat. Conversational interfaces bury change penalties, basic-economy restrictions, and baggage costs. Tap or click into the full fare rules and screenshot them. The price quoted in chat is rarely the price you’ll pay if anything goes wrong.
- Reset your profile before a major search. Default preferences saved from past trips can lock you into pricier cabins or inflexible dates. Log into your OTA or airline account, review marketing consent and personalization settings, then run a clean search in a private browser or a different app for comparison. Air Traveler Club’s tracking of temporary fare drops is one way to cross-check whether the price you’re seeing is genuinely competitive.
- Audit your data-sharing settings now. Before your next major search, review what your OTA and airline accounts are allowed to share with partners. In the EU, GDPR gives you limited opt-out rights on certain personalization features — use them if you want cleaner, less profiled pricing.
- Know when AI support will fail you. Automated disruption handling works for simple single-carrier rebookings. For multi-carrier itineraries, international connections, or anything involving a codeshare, get to a human agent faster than the chatbot wants you to. The automated message saying your trip is “being managed” is not the same as it actually being fixed.
Watch: EU AI Act enforcement guidance expected in 2026–2027 will determine whether European travelers gain meaningful new rights over how AI tools profile and price against them — or whether the current approach continues unchallenged.
Questions? Answers.
Is an AI trip planner actually cheaper than booking directly with an airline?
Not reliably. AI planners surface options from OTA inventory, which may include fees and markups not visible in the conversational interface. Airline direct sites often have the lowest base fare, and some carriers offer exclusive perks — seat selection, free changes — only when booked direct. Use AI to plan, then verify the total cost on the airline’s own site before paying.
What data does Booking.com’s AI Trip Planner collect about me?
The system logs search queries, clicks, device data, and purchase history to build a preference profile. This data is stored by Booking Holdings and may be shared with partner brands including Priceline and Kayak for marketing and cross-selling. EU users have disclosure rights and limited opt-out options under GDPR. Non-EU users have fewer formal protections and should review account privacy settings manually.
What happens if an AI-managed trip goes wrong mid-journey?
For single-carrier, simple itineraries, automated rebooking works reasonably well. For multi-carrier trips, international connections, or codeshare bookings, traveler reports consistently show that automated systems fail to resolve the full disruption — particularly when the problem involves two separate airlines. In those cases, call the operating carrier directly rather than waiting for the OTA chatbot to act.
Does the EU AI Act apply to travel booking platforms?
Yes, in part. The EU AI Act, formally adopted in 2024, classifies certain AI systems used for consumer profiling and personalized recommendations as potentially high-risk if they significantly affect consumer rights. Travel platforms operating in the EU face transparency, risk-management, and data-governance obligations. Enforcement guidance clarifying exactly how these rules apply to OTAs and airline apps is expected in 2026–2027.
