Expedia Group reported its third quarter on 4 November 2026, and the most useful line for hotel tech vendors was not a revenue figure. Asked what its deals with Meta's Muse and Google's AI products will cost, CEO Ariane Gorin said "the economics of the partnerships are being worked out", according to the earnings call transcript and Skift. The largest hotel distributors are now treating AI assistants the way they treated search engines twenty years ago: as acquisition channels with a price that has not been set yet.

This explainer covers what Expedia said, what each point means, and what it changes for how hotel tech vendors sell, price and position. It sits alongside our breakdown of Booking Holdings' Q3 AI numbers from the day before; read together, the two calls show how both large OTA groups are thinking about AI distribution.

What Expedia reported

The headline numbers, as reported by PhocusWire, MarketBeat and PYMNTS:

Gorin also said two thirds of Expedia's bookings now come direct, and CFO Derek Andersen said the number of B2B partners recently reached 75,000, per the transcript.

What "the economics are being worked out" means

Expedia confirmed it partnered with Google in AI Mode and that its Meta Muse integration was going live during the call. PhocusWire quotes Gorin saying Muse "is going to have access to our prices, our content and the like" and can either complete the booking itself or send travellers to Expedia.

On cost, Gorin's position was that Expedia has "a track record of being able to work with big advertising platforms or tech platforms to use them as channels to acquire customers at attractive economics" (Skift). In plain terms: Expedia expects to pay these platforms, and it expects to negotiate well because it brings scale, supply and servicing.

She also said agentic traffic is "very small", high intent, and not affecting direct traffic. "I think it's either incremental or it's taken from other paid channels. But right now, again, it's so small that it's hard to draw any conclusions," she said on the call.

For a vendor, the point is that the people with the most data are not claiming AI agents are a large booking channel today. They are claiming it will become a paid channel, and they are positioning to buy it cheaply.

Why Expedia will accept lower returns on OpenAI ads

Asked about OpenAI's new advertising products, Gorin said Expedia is participating and is willing "to maybe accept some returns that aren't as high as more established products that are more optimized" (PhocusWire). She added: "To the extent that there's more fragmentation at the top of the funnel, that's a good thing for us."

That last sentence matters most. A fragmented top of funnel favours whoever can afford to test every new surface at once. A large OTA can run test budgets on Google AI Mode, Muse and ChatGPT ads at the same time. An independent hotel cannot, and nor can most of the vendors selling to it.

Where the growth came from: B2B

The fastest growing part of Expedia was not its consumer brands. B2B bookings grew 19%, with double digit growth in all regions, and Gorin described the strategy as being "the one-stop travel platform for partners" (MarketBeat). In the quarter Expedia closed its CarTrawler acquisition, launched an activities API, and made Qantas the first partner on its new agentic template, which combines Expedia's B2B technology, supply and servicing with a customisable interface.

Gorin also said that if AI companies decide to sell travel themselves rather than act as a demand channel, Expedia is "clearly well positioned" to supply them (PhocusWire). So Expedia is selling to the AI platforms as well as buying from them.

AI inside the booking flow

Expedia listed several AI features already live, according to PYMNTS and the transcript:

• An AI compare agent and a property expert agent in the lodging booking flow, which help travellers weigh hotels and amenities without opening each property page. Gorin said this leads to more conversion and larger purchases.

• An agentic voice experience on Vrbo that resolved over 15% of inbound contacts in its first two months.

• A post-booking activity agent; travellers who used it were twice as likely to book an activity within seven days.

• An autonomous hotel onboarding agent, built in less than a quarter by a small team, work that Gorin said would once have taken multiple quarters and many more people.

Note what the property expert agent does: it answers questions about a hotel using the content Expedia holds. Whatever the hotel has loaded into the OTA becomes the answer. We covered this pattern in how OTA AI answers property questions.

What this means for how hotel tech vendors sell

Booking engines and direct booking tools

Expect hoteliers to hear that AI traffic is tiny and conclude they can wait. The better argument is about who will own the connection when it grows. Some vendors are already building direct routes: Hospitality Investor reports that Cloudbeds launched a direct integration with Meta Muse in October 2026, so bookings land in the hotel's own system as direct bookings. If you sell direct booking, your pitch is now "be reachable by agents without an intermediary", not "AI is replacing OTAs". The Google AI Mode booking changes make the same point.

Channel managers and connectivity

AI platforms need prices, availability and content in a structured form, and Expedia is pitching itself as the supplier of exactly that. Connectivity vendors that can show clean, fast rate and content delivery to new endpoints have a story worth telling to hotel groups now, before the OTAs make it the default route.

Revenue management and commercial tools

If AI distribution becomes a paid channel, hotels will need to know what each booking costs net of commission and acquisition cost, not just the gross rate. Clanalytix founder Patrick Angwin told Hospitality Investor that owners should report "net rate and contribution by channel rather than revenue alone". Tools that show channel contribution can use that line directly. For the commission argument itself, see using OTA commission pressure in a sales argument.

Onboarding and services

An OTA building an autonomous hotel onboarding agent in under a quarter sets a new reference point for how long it should take to get a property live. Hoteliers will compare. If your onboarding still takes weeks of calls, shortening time to first value is now a competitive issue, not just a retention one.

What not to take from the call

• Do not claim AI agents are already a big booking channel. Expedia itself says the volume is too small to draw conclusions.

• Do not tell hoteliers OTAs are losing to AI. Expedia grew room nights 7% and says two thirds of its bookings come direct to its own brands.

• Do not price an AI feature on booking volume you cannot yet show. Price on what you can measure today, such as content quality or contact deflection, and check your plan against the hotel software Pricing Index.

The useful question for your next quarter is where your product sits as AI distribution turns into a paid channel: helping hotels reach agents directly, helping them measure what agent bookings cost, or competing with the OTAs' own tools. That is a positioning decision, and it belongs in the position and price layers of your hotel tech go-to-market strategy.

If you want to know whether your current pitch still holds after this earnings season, book a free GTM leak audit and we will look at where it is losing deals.

Frequently Asked Questions