Forget the chatbot hype. Airbnb’s real AI story is happening behind the scenes, and it’s altering the economics of how the company builds software. Co-founder and CEO Brian Chesky dropped concrete numbers on the latest earnings call: AI now generates 60% of the company’s code. More startling is the velocity that unlocks. Chesky pegged the reduction in time from concept to product launch at 60%, while the total volume of features shipped jumped nearly 80% compared to the same six-month period last year.
That’s not marginal efficiency. It’s a structural shift in output that touches core transactional plumbing — search, sign-up, checkout, and payments — not just cosmetic interface tweaks. Host onboarding flows have also been rebuilt with AI assistance, compressing what was likely a friction-heavy process.
On the consumer side, Chesky is finally budging from his long-held skepticism of slapping a generic chatbot onto a travel app. Airbnb will begin testing an AI-powered search experience, but with a notable design choice: it’s opt-in. Users keep their familiar filters and maps, and a toggle lets them switch to natural language queries that return conversational, AI-generated titles in a visual format. Highlights on product description pages get personalized in real time.
Customer support is where the AI deployment is deepest. The company’s AI agent, live in North America since early 2025 and now in over 50 languages, resolves nearly 45% of issues without a human touch. That’s driven support costs per booking down 16% year-over-year. Voice call support is slated for later this year. The financial backdrop matters here: Airbnb posted $3.6 billion in quarterly revenue, up 17%, with adjusted EBITDA surging 21% to $1.3 billion. When a platform at that scale can autonomously close nearly half its support tickets while simultaneously accelerating product development by 80%, you’re looking at a legitimate compounding advantage — not just a press release statistic.
