Airbnb plans to bake in AI features for search, discovery and support

Airbnb has been slow and deliberate in rolling out AI within its app. Still, CEO Brian Chesky said Friday that the company is now preparing to integrate new features powered by large language models. The goal is to help users search for listings, improve discovery, plan trips more easily, and assist hosts with managing their properties and operations.

During Airbnb’s fourth-quarter conference call, Chesky said the company wants to expand the role of large language models across customer discovery, customer support, and engineering workflows.

“We are building an AI-native experience where the app does not just search for you. It knows you. It will help guests plan their entire trip, help hosts better run their businesses, and help the company operate more efficiently at scale,” Chesky said.

Separately, Airbnb said it is testing a new tool that lets users search for properties and ask questions about listings and locations using natural language.

Currently, Airbnb offers an LLM-powered customer service bot, some personalisation features, and messaging tools. The company said the new AI search feature is expected to “evolve into a more comprehensive and intuitive search experience that extends through the trip.”

When asked by an analyst whether Airbnb intends to include sponsored property placements inside AI search, Chesky said the company first wants to get the design and overall user experience right.

“AI search is live to a very small percentage of traffic right now. We are conducting extensive experimentation. Over time, we are going to be experimenting with making AI search more conversational, integrating it into more than the trip, and, eventually, we will be looking at sponsor listings as a result of that,” Chesky said, adding that Airbnb would consider an advertising format that fits naturally into a conversational search flow.

Chesky said Airbnb plans to lean on the AI experience of its newly hired CTO, Ahmad Al-Dahle — who previously worked on Meta’s Llama models — and combine that technical expertise with Airbnb’s large store of identity and review data to make the app more useful.

Airbnb also said its AI-powered customer support bot, which launched in North America last year, is now resolving a third of customer issues without any help from human agents. Chesky added that Airbnb plans to enable customers to call the AI bot for support and to broaden language coverage.

“A year from now, if we are successful, significantly more than 30% of tickets will be handled by a custom service agent, in many more languages, in all the languages where we have live agents. AI customer service will not only be chat, but it will also be voice,” he said.

Internally, Airbnb is also looking to increase AI adoption. The company said 80% of its engineers currently use AI tools, and it aims to reach 100%. Airbnb reported quarterly revenue of $2.78 billion in the fourth quarter, which came in above expectations and represented a 12% increase compared with the same period a year earlier.