Airbnb's Chesky Built an AI Brain That Cut His Meetings in Half

Brian Chesky says he has cut his meetings in half. The Airbnb co-founder and CEO built an AI "brain" trained on 18,000 Keynote presentations, 50,000 Google Docs and other company materials, and it now serves as his primary window into what is happening inside the company.

Chesky laid out the details earlier this week at the Skift Global Forum. His diagnosis of the problem was blunt: "most of us CEOs of big companies don't really know what is going on in our companies."

"So we have these things called meetings, which are just for us to figure out what's going on, and well-intentioned people underneath us kind of manage us," he said, per Business Insider.

The workaround became a personal engineering project. Chesky described his core objective in one sentence: "My whole job was: how can I not need meetings to know what's going on inside the company?"

The answer is what he calls a "brain," or an "intelligence corpus," fed with the company's own internal record. He paired it with a suite of dashboards built to give him a clearer, real-time view of operations across Airbnb. The result, in his telling, is near-total visibility without the calendar burden.

"This intelligence layer means I now have total information," he said. "I've cut my meetings in half."

Chesky stressed that the finished product is for himself, not for customers. It is a management instrument, not a product roadmap item. But it reflects a broader transformation he is pushing through the entire company.

An "AI-pilled" executive team

At the forum, Chesky said Airbnb is now "pretty much" AI-centered and that his executive team has been "AI-pilled" — his shorthand for an internal shift toward adopting the technology across the business.

He has also made his own fluency a personal priority. Chesky said he wants to be as adept with AI as someone who grew up with it. "I never wanted to be this person that it passed me by," he said.

That stance places him at the aggressive end of a corporate leadership cohort that remains divided on generative AI. KPMG's 2026 survey of 100 CEOs at large U.S. companies found that 77% believe generative AI may have been overhyped during the previous year.

The skepticism has not slowed the spending. Nearly 80% of those surveyed CEOs said they plan to devote at least 5% of their 2026 capital budgets to AI initiatives. Another 41% said AI will account for 10% or more of that spending.

The numbers behind the shift

At Airbnb, Chesky attributes concrete operational results to the internal adoption push. The company has shipped more than 600 features on the Airbnb site this year, he said, including AI-powered customer support and AI search. That figure is nearly double last year's total, and he credited part of the jump to the executive team's use of AI tools.

The customer-support function shows the clearest cost effect. AI now handles close to half of Airbnb's customer-service tickets across 50 languages, according to Chesky. That has helped the company reduce support costs per booking by 16% year over year.

Those are the kind of numbers that turn a management philosophy into a margin story. A 16% reduction in support cost per booking, applied across a business of Airbnb's transaction volume, is a material line item rather than a laboratory experiment.

The 2027 test

The open question is whether Airbnb can convert internal efficiency into a customer experience that visibly differs from what rivals offer. Chesky has put a date on it: the company will launch its own AI agent in 2027, with an interface designed to be "much richer than a chatbot."

Until then, the internal evidence is the story. A CEO who replaced half his meeting load with a corpus of company knowledge, an executive team running AI tools by default, and support costs falling 16% per booking together suggest the shift is already embedded in how Airbnb operates — not a pilot awaiting approval.