Booking CFO: Even AI hyperscalers don't know ROI

Steenbergen's remarks came at the Fortune AIQ Summit at the New York Stock Exchange, in a conversation with Fortune senior writer Sheryl Estrada. He framed the past two years as a period of experimentation and said the next phase means rebuilding processes from a white sheet. Customer service went first, and the company has since pointed to early signs in its Q2 numbers: connected trip transactions grew at a low-double-digit rate at Booking.com, merchant bookings reached about 73% of gross bookings, and Level 2 and Level 3 Genius members made up more than 30% of active customers. On the customer side, he said on the company's Q2 call in August that referrals from large language models, paid and unpaid, were under 1% of total room nights.

The company's public pitch is that AI will make travel easier for customers and its marketing cheaper. Steenbergen said AI won't change how travel fundamentally happens, only the experience, calling it the connected trip. Travelers visit about five platforms on average before booking, and about one-third of Booking's customers arrive through paid channels that cost the company $8 billion to $9 billion a year. He said the bigger opportunity is growth rather than cost, with AI potentially making Booking a much more high-frequency business through proactive suggestions when plans change.

Steenbergen also described his own setup: an AI coach, plus two agents, one a strategic thought partner for board presentations and one a critical equity research agent for earnings prep. The reading here is that Booking's AI story is still mostly about internal efficiency, and the customer-facing gains he cited remain modest, with limited data. Whether the growth case materializes may depend on how quickly the connected trip moves from early signals to a larger share of transactions.