Airbnb has been cautious about slapping AI onto its consumer product. But behind the scenes, the company has been moving fast. According to TechCrunch, CEO Brian Chesky told investors during Airbnb’s second-quarter earnings call that AI has cut the time from concept to launch by as much as 60%, and that the company shipped nearly 80% more features in the first half of this year compared to the same period in 2024. Earlier in the year, Airbnb had already disclosed that AI is writing 60% of its code. The numbers are starting to add up.
This matters because Airbnb’s approach to AI has been notably different from companies like Google, Booking.com, or Expedia, all of which have moved more aggressively to put AI-generated results and chat interfaces directly in front of users. Chesky has publicly pushed back on the chatbot model for travel, arguing that a conversational interface alone doesn’t map well to how people actually book trips. So Airbnb has been channeling its AI investment internally, into engineering velocity and support infrastructure, while keeping the product surface mostly familiar.
That’s starting to change. Chesky confirmed the company is now testing an AI-powered search feature. But Airbnb is doing it carefully. Rather than replacing the existing search and filter experience that users already know, the company is adding a toggle. Users who want to try it can switch to AI search and type in natural language queries. Results will be returned in a visual format, with AI-generated titles that Chesky described as conversational in tone. Listing highlight descriptions on the product page will also be AI-generated in real time and personalized to the individual user.
It’s a measured rollout, and that’s probably smart. Forcing an unfamiliar interface on users who are comfortable with the current product is a real risk, especially in a category like travel where people can be anxious about bookings and tend to prefer clarity over novelty. The toggle approach lets Airbnb gather data and iterate without disrupting the core experience.
The more concrete wins right now are on the support side. Airbnb launched its AI support agent in North America in 2025 and has since expanded it to more than 50 languages. Voice call support is in the pipeline for later this year. The company says that 45% of customer issues handled by the AI agent are resolved without any human involvement. Support cost per booking is down 16% year over year. Those are real operational savings at scale.
AI has also touched areas like search ranking, sign-up flows, checkout, and payments, along with host onboarding. The productivity gains across these areas are what’s driving the 80% increase in feature output. This is the part that often gets overlooked in the consumer AI narrative: the biggest near-term value for many companies isn’t a new product feature users can see, it’s faster internal development cycles.
Airbnb’s financials for the quarter ended June were strong. Revenue came in at $3.6 billion, up 17% year over year. Adjusted EBITDA hit $1.3 billion, a 21% increase. The company isn’t crediting AI for all of that, but the connection between faster shipping, lower support costs, and margin expansion is hard to ignore.
For developers and product teams watching this space, Airbnb is a useful case study in sequencing. Build AI into the development process first, prove the economics on support, and only then introduce AI-facing features to users, with an opt-in rather than a forced migration. It’s not the flashiest approach, but it’s one that’s showing measurable results.
