Airbnb's AI assistant resolves 40% of customer inquiries

Title

Airbnb's AI assistant resolves 40% of customer inquiries

Content

Airbnb's AI customer service assistant resolved over 40% of customer inquiries in Q1 2026, up from resolving about one-third of customer questions in the prior quarter, co-founder and CEO Brian Chesky said on the Q1 2026 earnings call. Resolution time "significantly" improved, and the company saw cost per booking decrease about 10% year over year in Q1. "We've seen the cost per booking decrease about 10% year over year in Q1, and we expect to see more of this as we improve AI customer support this year," Chesky said. While many online travel agencies decided to implement AI in the top of the funnel with AI-powered search, Airbnb took a different strategy and started at the bottom of the funnel. "The reason we decided to start with this is we want to focus on the hardest problem in AI, which we thought was customer service," Chesky said. "The reason why is the stakes are high, you cannot hallucinate, you have to answer things very, very quickly because they are calling and they have problems." Airbnb rolled out its AI customer service assistant in April 2025. The technical challenge is significant: the AI assistant must be multilingual (hosts and guests often don't speak the same language), adjudicate difficult issues, and know when to pass the call to a live agent. "You have to escalate to a human accurately, especially if it's timely or there's a trust and safety incident," Chesky said. The system must also handle personally identifiable information (PII) and protect data, read and train on nearly 100 internal policies, tens of thousands of evolving conversations, and millions of data points of how prior cases were adjudicated to answer correctly. Operationally, Airbnb's AI assistant has scaled to a multi-million-user customer service base, with the company processing high-volume inquiries daily across multiple languages. The system integrates with Airbnb's existing case management and escalation infrastructure to ensure smooth handoff to human agents when needed. Chesky emphasized that AI is not magic: "when you bring AI under the hood, you realize that in order to be good at AI, you need to be really good at technology, foundational. You need to have really good data and infrastructure." Airbnb has been preparing its data warehouse for the past few years, following the mantra that "your AI is only as good as your data." Airbnb has also deployed AI in the listing-viewing phase. The company plans to roll out more AI features in the mid-funnel consideration phase on May 20. "One of the things our guests told us is when they get to an Airbnb, it's great when they see like 100 reviews, it's awesome, but they don't have time to read all 100 reviews. So we now have AI summaries." The company is also experimenting with deploying AI at the top of the funnel with AI search. Future direction: pilot AI in search across multiple touch points — post-search, mid-search interruption for suggestions, filter panels, and post-booking.