Airbnb disclosed that AI has cut its feature development cycle by 60%, from concept to launch, and increased the number of features shipped by nearly 80% compared to the prior year.
The company is testing a new AI-powered search function that lets users type natural language queries instead of using traditional filters, though users can toggle back to the old interface.
Airbnb's AI customer support agent, live in North America since 2025, now operates in more than 50 languages and resolves 45% of issues without human help, reducing support costs per booking by 16%.
What happened
Airbnb CEO Brian Chesky said on the company's second-quarter earnings call that AI has reduced the time from concept to launch by as much as 60%, and the company shipped nearly 80% more features and improvements in the first half of 2025 compared to the same six months last year. Earlier this year, Airbnb disclosed that AI is writing 60% of its code. The company is now testing a new AI search feature that lets users toggle to natural language search with AI-generated titles and personalized highlights.
Why it matters
Airbnb's internal AI adoption is speeding up product velocity across core areas—search, sign-up, checkout, and payments—while keeping consumer-facing rollout measured. On the support side, the company's AI agent launched in North America in 2025 and now operates in more than 50 languages, resolving nearly 45% of customer issues without human intervention and cutting support cost per booking by 16% year-over-year. This signals how AI is reshaping both how travel companies build and how they serve customers.
What to watch
Airbnb is testing the AI search toggle with an opt-in design, allowing users to stay with the traditional search-and-filter interface if they prefer. The company plans to make its AI customer support bot available for voice calls later this year. For context, Airbnb posted revenue of $3.6 billion (up 17% year-over-year) and adjusted EBITDA of $1.3 billion (up 21%) for the quarter ended in June.
During its second-quarter earnings call, Airbnb disclosed that AI has become central to its product development process. CEO Brian Chesky stated that the company has reduced time from concept to launch by as much as 60% and shipped nearly 80% more features and improvements in the first half of 2025 compared to the same period last year. Earlier in the year, Airbnb revealed that AI is writing 60% of its code, indicating the depth of the company's reliance on AI-assisted development across teams.
On the consumer side, Airbnb is testing a new AI search feature designed to let users interact with the platform more naturally. Rather than clicking filters and browsing listings, users can type requests in natural language and receive results in a visual format. The listing titles and highlights are AI-generated and personalized to each user. Chesky described the experience as conversational, "as if you're reading a chatbot, but more visual." Importantly, Airbnb is implementing an optional toggle so users can switch between the new AI search and the traditional search-and-filter interface—a deliberate choice to avoid forcing change on users accustomed to the current experience. Chesky has maintained that a chatbot-like interface alone would not suit travel use cases, so the company has focused on AI for search, discovery, and support instead.
Airbnb has deployed AI most aggressively in customer support. The company launched its AI-powered support bot in North America in 2025 and has since expanded it to more than 50 languages, with plans to enable voice calls later in 2025. The results have been striking: nearly 45% of customer issues that begin with the AI agent are resolved without human intervention. This efficiency gain has translated directly to the bottom line—support cost per booking is down 16% year-over-year. Chesky also noted that AI has bolstered areas like checkout and payments, and helped hosts with faster onboarding flows.
For the quarter ended in June, Airbnb posted revenue of $3.6 billion, up 17% year-over-year, and adjusted EBITDA of $1.3 billion, up 21% year-over-year, demonstrating that the company's financial momentum remains strong despite the ongoing refinement of its AI strategy.
Airbnb's earnings call revealed a widening gap between its internal adoption of AI and its consumer-facing rollout. While AI has become foundational to how the company builds—writing 60% of its code and cutting development cycles by 60%—the company remains cautious about imposing AI interfaces directly on users who are accustomed to traditional search and booking flows. This tension reflects a broader challenge in the travel industry: AI can accelerate product teams, but the end user may not be ready to abandon familiar workflows. By offering the AI search toggle, Airbnb is hedging: it gets to test natural-language discovery without forcing change on the majority.
The most tangible impact so far is in customer support. Airbnb's AI agent, now live in more than 50 languages, resolves nearly 45% of incoming issues without escalation to humans—a scale of automation that has directly lowered support cost per booking by 16% year-over-year. This suggests that AI's highest ROI for Airbnb may lie not in serendipitous discovery (the romantic ideal of "surprise me") but in handling the high-volume, repetitive work that traditionally drains operational budgets. The company's Q2 results—$3.6 billion revenue (up 17%) and $1.3 billion adjusted EBITDA (up 21%)—suggest the business remains healthy, giving Airbnb room to experiment with consumer AI without the pressure to accelerate monetization.
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytime
Ask AI anything about this article. Q&As are published on this page for other readers too.
QumulusAI announced a GPU-as-a-Service agreement with DRW, a global trading firm, to supply a dedicated NVIDIA…
HireRoad, an HR software company, scrapped its 18-month legacy product rewrite plan and instead reorganized it…
OpenAI introduced Premium Seats for ChatGPT Business, priced at $125 per user per month ($100 with annual bill…
Researchers at A Security disclosed vulnerabilities in Zoom's screen-sharing annotation protocol on Tuesday th…
Anthropic pledged to embed machine-readable watermarks in Claude-generated text and digitally signed provenanc…
OpenAI announced that its unreleased model Astra had produced solutions to 10 long-standing mathematics proble…
The AI news that matters, in one minute each morning.
Sign up free