OpenAI's ChatGPT travel app integrations with Expedia and Booking.com, announced at DevDay on October 6, 2025, have been live for roughly ten months as of April 2026. The headline conclusion after a full booking cycle: ChatGPT is genuinely useful for the search-and-compare phase of a trip, and it still hands you off to a partner site to actually book. That handoff is where the points-and-miles tradeoffs kick in.
OpenAI confirmed the integrations through its Apps SDK announcement at DevDay 2025, with Expedia, Booking.com, Spotify, Zillow, and Canva named as launch partners. Tripadvisor, Uber, and OpenTable's TheFork have since rolled in, according to OpenAI's developer changelog and partner press releases through the first quarter of 2026. The Apps SDK is built on the Model Context Protocol, the same MCP standard now used across multiple AI platforms.
What ten months of usage has clarified
The pattern that has emerged is consistent across the travel-press coverage and our own testing: ChatGPT is a strong front-end for itinerary research and a weak front-end for the actual transaction. You can ask it for flights from Chicago to Lisbon under $700 in shoulder season, and Expedia returns live results in the chat. You can ask Booking.com for boutique hotels in Porto under $250 per night with a pool, and the listings appear with photos, ratings, and availability. The conversational flow, asking follow-ups about neighborhoods, walking distances, or breakfast inclusions, is meaningfully faster than juggling browser tabs.
The handoff is where the model's value drops off. When you click through to book, you land on Expedia or Booking.com to complete the transaction. That is fine for cash-rate hotels and basic flights. It is a problem the moment loyalty matters. ChatGPT cannot log you into your United MileagePlus account and pull award seats. It cannot search Hyatt Globalist availability. It cannot confirm whether a Hilton property will recognize your Diamond status. Those flows still require direct booking.
Where the integration genuinely earns its place
For the planning phase, ChatGPT replaces three or four browser tabs with one conversation. Travelers comparing dates across a flexible window, mapping a multi-stop European itinerary, or surfacing destinations that fit a fuzzy budget brief get real time savings. The AI handles vague queries like "warm beach for a family of four in February under $3,000 all-in" better than dropdown filters do.
It is also useful for sanity-checking pricing. If a hotel quotes $340 a night on its own site, asking ChatGPT to surface Expedia and Booking.com rates in the same conversation gives you a quick read on whether you are seeing a fair price. That is a research function, not a booking one.
Where direct booking still wins, and why
For airline tickets, direct booking remains the default for anyone collecting miles. Award redemptions, upgrade lists, mixed-cabin itineraries, and elite-status benefits like preferred seats, free changes, and mileage accrual at full rate are tied to the airline's own system. Expedia and Booking.com bookings often earn reduced or zero loyalty credit, depending on the carrier's current policy, and changes routed through an OTA add an extra layer of friction when something goes wrong.
For hotels, the same logic holds with sharper edges. Marriott Bonvoy, Hilton Honors, World of Hyatt, and IHG One Rewards all reserve their strongest perks (points earning, late checkout, room upgrades, breakfast at certain tiers) for direct bookings. Booking through an OTA generally voids those perks even at properties where you hold elite status.
Premium travel cards add another layer. The Chase Sapphire Reserve, American Express Platinum, and Capital One Venture X all extend their strongest trip-protection coverage to charges made directly with the airline, hotel, or rental car company. OTA bookings still earn rewards but can complicate claims for delays, cancellations, and lost bags.
How to actually use this in 2026
The hybrid workflow that has settled out across the points-and-miles community looks roughly like this. Use ChatGPT for the messy front end: brainstorming destinations, comparing price across flexible dates, narrowing a hotel shortlist. Once you have a target, meaning a specific flight, a specific property, or a specific date range, switch to the airline or hotel's own site to book. If you are paying cash for a flexible-rate Hilton, OTA versus direct is mostly a wash on price. If you are using points, chasing status, or expecting to use card travel protections, direct is the only sensible channel.
Privacy is also worth flagging. OpenAI's Apps SDK requires partner apps to follow data-minimization rules and request explicit consent before connecting an account, per OpenAI's October 2025 documentation. Granular per-category data controls were added in early 2026. Adding ChatGPT to the chain still means one more party touches your travel data, which matters more for some readers than others.
What the integration signals about loyalty's future
The deeper story is what AI mediation does to brand differentiation. If a generation of travelers gets used to asking ChatGPT for "the best hotel in Lisbon under $250," the brands they end up with will be the ones whose loyalty programs make the strongest case for booking direct after the AI surfaces the option. Programs offering meaningful elite benefits, generous earning, and clear cash-equivalent value on points are the ones positioned to keep customers in their direct channels. Programs that have been gradually devaluing benefits are the ones most exposed to AI commoditization.
For now, the practical takeaway is unchanged from a year ago: ChatGPT is a research tool, direct is the booking channel, and your loyalty stack is still where the real value sits.
This article contains affiliate links. If you apply through our links, we may earn a commission at no cost to you, which helps us continue sharing points and miles strategies with the community.
Some of the links in this article are affiliate links. We may receive a small commission at no extra cost to you if you apply through these links. This helps us keep the site running and continue creating free content.
