Airbnb reports significant gains in product velocity through internal automation
The company has reduced the time required to move features from conceptualization to deployment by 60 percent as it prepares to test natural language search.
Julian Reeve
Aug 7, 2026 · 1 min read
Sixty percent of the code currently being deployed by Airbnb is generated by artificial intelligence. During a second-quarter earnings call, chief executive Brian Chesky disclosed that the company has shortened the duration between feature conceptualization and launch by the same percentage. The shift represents a fundamental change in the company's internal operations, resulting in an 80 percent increase in the volume of features and improvements shipped compared to the same period last year.
While the company has focused heavily on backend efficiencies, its consumer-facing implementation has remained measured. Airbnb is now moving to test a natural language search interface, though it will maintain a toggle to preserve the existing filter-based system for users who prefer it. The proposed interface will generate visual results and conversational product descriptions in real-time, personalized to individual user profiles.
On the operational side, the deployment of automated support agents has begun to alter the company's cost structure. A multilingual bot now handles nearly 45 percent of customer issues without human intervention, contributing to a 16 percent year-over-year decline in support costs per booking. These efficiencies arrived alongside a 17 percent increase in quarterly revenue, which reached $3.6 billion.