On Airbnb’s November 5, 2026, third-quarter earnings call, CEO Brian Chesky described bookings from Meta’s Muse and Instinct as negligible compared with Airbnb’s daily volume. The company’s response is to build a travel-specific AI agent of its own, with a rollout planned for the first half of 2027.
14
10
What Chesky said about Muse and Instinct
“We’re not seeing a lot of traffic, and even fewer bookings from Muse and Instinct yet,” Chesky said. He put Airbnb’s overall volume at nearly half a million bookings a day and called the number coming through those agents “very non-material.”
14
He also said traffic from the agents had peaked in early October and had not increased over the previous four weeks, remaining stable during that period. That describes the activity Airbnb was seeing on the call; it does not establish how agent traffic might develop later.
14
Why Airbnb is building for travel
Chesky contrasted general-purpose agents’ performance on routine tasks—such as restaurant reservations or scheduling appointments—with the more involved task of planning a trip.
10 His argument for a specialized product is that a travel agent should be designed around travel decisions rather than treat them like a generic chat exchange. Reporting on the call quoted him saying a specialized agent would need to be substantially better at its task than a general-purpose alternative to overcome user inertia and switching costs.
30
Airbnb said its agent would build an interface around a traveler’s question instead of replying only in a conventional chatbot window. The planned components include maps, photos, comparisons, calendars and itineraries, assembled to suit the request.
10
19 The company has also described an ambition to connect specialized agents with partners and support bookings across travel categories.
7
The planned launch and integrations
Airbnb said it expects to begin rolling out its own AI travel agent in the first half of 2027.
10 On the call, Chesky also discussed integrating multiple MCPs with the trip agent in that period.
4 The available reporting supports that planned integration work, but does not establish a timetable for opening an API to outside agents or explain how sponsored listings would appear in the experience.
The wider AI and earnings context
Airbnb reported third-quarter revenue of $4.9 billion, up 19% year over year; gross booking value rose 21% to $27.7 billion, and its combined “nights and seats booked” metric increased 16% to 155 million. The combined figure is not a room-night-only measure.
33 The company also raised its full-year revenue-growth outlook to the “high teens” and authorized a $12 billion share buyback.
8
For comparison, Skift reported that Expedia’s third-quarter revenue grew 12%, while Booking Holdings’ revenue fell 1%; the report attributed Booking’s decline in part to a $1 billion accounting hit related to a possible FTC settlement. These financial results describe a competitive backdrop, not evidence that Airbnb’s performance was caused by its AI strategy.
25
Airbnb’s customer-support AI is a separate part of that strategy. Reporting on the company’s third-quarter results said its assistant resolved about 40% of support issues without a human, while support cost per booking fell more than 17% year over year.
11 Separately, Airbnb disclosed that an AI lab controlled by Chesky paid the company $7.1 million for intellectual property and certain expenses, and that Airbnb could invest in the lab. The provided reporting does not establish a direct connection between that arrangement and the planned travel agent.
5
The clearest takeaway from the call is therefore narrow but useful: third-party agent bookings were tiny relative to Airbnb’s current volume, while the company was preparing a purpose-built travel experience. Chesky’s remarks describe present traffic and Airbnb’s product direction—not a settled verdict on the future importance of AI-driven bookings.