Airbnb is moving faster with AI behind the scenes than it is in the parts of the app that travelers see every day. The company says the technology is already changing how quickly it builds, tests, and ships product updates.
On its second-quarter earnings call, co-founder and CEO Brian Chesky described AI as a force multiplier across Airbnb's product work. The clearest signal is speed: Airbnb says some key initiatives now move from concept to launch as much as 60% faster.
AI is changing how Airbnb builds product
Airbnb has been cautious about putting consumer-facing AI at the center of the travel booking experience. But internally, the company has embraced it more aggressively.
Earlier this year, Airbnb said AI is writing 60% of its code. In the latest update, Chesky said AI is also helping the company increase the pace of product development beyond code generation alone.
He said Airbnb is now building, testing, and iterating faster than it could a year ago. Compared with the same six months last year, the company has increased the number of features and improvements shipped this year by nearly 80%.
That matters because Airbnb's product touches several high-friction areas of travel. Chesky pointed to search, sign-up, checkout, and payments as areas where AI has helped the company move faster. Airbnb has also released features intended to help hosts, including a quicker onboarding flow.
The new AI search test keeps the old interface in place
The most visible upcoming change is AI search. Airbnb plans to start testing a version of search that lets users type requests in natural language and receive results in a visual format.
The company is not forcing that experience on users who are familiar with the existing search and filter flow. Instead, Airbnb is adding a toggle so customers can switch into AI search when they want to try it.
That approach reflects a broader product bet. Chesky has said that a chatbot-like interface alone does not fit travel use cases. Airbnb has focused instead on AI for search, discovery, and support, where the technology can help interpret intent while still keeping the browsing experience visual.
In the planned AI search experience, the response can include AI-generated titles that feel conversational. From there, a user can move to the product description page, where listing highlights can be generated in real time and personalized to that user.
The key product question is how much AI should guide discovery without making the experience feel unfamiliar. Airbnb's answer, at least for this test, is to keep the current search and filters available while offering AI search as an option.
Consumer AI has arrived slowly
Airbnb's consumer-facing AI features have been limited so far. The company has used AI for review summaries and listing highlights, but it has not yet remade the main booking interface around an AI assistant.
That slower rollout is notable because travel search can be open-ended. People often know the kind of trip they want before they know the exact place, dates, or property. Natural language search could help users express those preferences more directly, but Airbnb is still testing how that should appear inside the product.
The company is also balancing habit. Many customers already understand how to use traditional filters for location, price, property type, amenities, and other criteria. A toggle lets Airbnb test a more conversational search layer without removing the familiar path.
Support automation is already producing measurable results
Airbnb has deployed AI more heavily in customer support than in the main consumer booking interface. The company launched its AI-powered bot in North America in 2025 and has since expanded it to more than 50 languages.
Airbnb also plans to make the bot available for voice calls later this year. That would move the support AI beyond text-based interactions and into another major customer service channel.
The company says nearly 45% of customer issues that begin with its AI agent are completed without human intervention. Airbnb also says its support cost per booking is down 16% year-over-year because of this automation.
Those figures show why support is a natural early home for AI at Airbnb. Customer issues often require fast triage, clear routing, and answers across many languages. If the AI agent can resolve a large share of those interactions on its own, the company can reduce support costs while keeping human agents available for cases that need them.
Airbnb ties AI progress to quarterly growth
The AI update came alongside positive quarterly results. For the quarter ended in June, Airbnb reported revenue up 17% year-over-year to $3.6 billion. Adjusted EBITDA rose 21% to $1.3 billion.
Those results do not mean AI alone caused the quarter's growth. The source information does not make that claim. What Airbnb is saying is narrower and more concrete: AI is helping the company ship more product work faster, and support automation is lowering support cost per booking.
For Airbnb, the near-term AI story is therefore split into two tracks. The first is operational, where AI is already embedded in coding, product development, and customer support. The second is customer-facing, where Airbnb is testing how natural language search can fit into a visual travel marketplace without replacing the search tools users already know.