Airbnb uses AI to write 60% of new code and accelerate product shipping

Airbnb is integrating artificial intelligence into its core user discovery experience while aggressively deploying the technology to overhaul its internal engineering workflows. The company is currently testing a new search experience that allows users to utilize natural language queries via a toggle switch, marking a significant evolution in how travelers interact with the platform.

According to reports from TechCrunch, the impact of AI on Airbnb's internal operations has been substantial. CEO Brian Chesky stated that the company has reduced the time from concept to launch for key initiatives by as much as 60%. This acceleration is driven by a deep integration of AI into the development pipeline, with AI now writing 60% of Airbnb's new code. As a result, the volume of features and improvements shipped over the last six months has increased by nearly 80% compared to the same period last year.

A shift in AI strategy

Historically, Airbnb has maintained a cautious approach toward consumer-facing AI. Brian Chesky previously argued that standard chatbot interfaces were unsuitable for the complexities of travel use cases. Until now, the company's public-facing AI efforts were largely confined to listing highlights and review summaries.

However, the internal efficiency gains have been undeniable. Beyond engineering, the company has seen a measurable impact on its bottom line through automated support. Nearly 45% of customer issues that begin with an AI agent are now completed without any human intervention, a shift that has reduced Airbnb's support cost per booking by 16% year-over-year.

Modernizing the discovery experience

This strategic pivot represents a move from using AI primarily for back-office efficiency to embedding it into the primary user journey. By implementing the new natural language search via a toggle, Airbnb is attempting to modernize its discovery process without alienating a user base that still relies on traditional filters. This hybrid approach allows the company to test the viability of generative search while maintaining the reliability of its existing interface.

The road ahead

As Airbnb continues to shorten its product development lifecycle, the industry will be watching to see if these internal speed gains translate into a superior user experience. While the internal metrics show a massive increase in shipping velocity, the success of the natural language search toggle will determine if Airbnb fully commits to an AI-first interface. For now, the company remains focused on the duality of using AI to build the product faster while carefully calibrating how that AI is presented to the traveler.