64% of Asia-Pacific Travelers Plan With AI. Under 1% Book Through It

A Booking.com survey earlier this year found 64% of Asia-Pacific travelers use AI to plan trips. Yet traffic referred directly by large language models (chatbots such as ChatGPT and Gemini) accounts for well under 1% of Booking Holdings' room nights, according to the parent company.

That gap is the whole story for Laura Houldsworth, Booking.com's vice president and managing director for Asia-Pacific. Speaking at the Skift Asia Forum in Thailand on April 29, she said the company is comfortable with travelers starting their research inside AI chatbots — as long as the booking still closes on Booking.com.

Why doesn't AI research become AI bookings?

Houldsworth attributes the mismatch to traveler behavior, not platform failure. "They might be searching on Claude one minute, ChatGPT another, Gemini, then rounding out with Google, and then coming to us directly," she told Skift. In that common journey, no single assistant earns referral credit, which makes LLM traffic look vanishingly small even as AI shapes demand.

Her argument is that trust decides where the money lands. Booking.com brings roughly 400 million reviews, secure payments, and customer service — things chatbots don't offer. AI may own the start of the trip; Booking.com wants to own the transaction.

How does distribution change for sellers?

The same logic governs Booking.com's approach to superapps (multipurpose messaging-and-payment platforms dominant in Asian markets). Daily usage of those apps hasn't translated into big-ticket travel spend — the company's experience suggests people chat there but don't book vacations there. So rather than build a rival superapp, Booking.com is plugging into the platforms travelers already use.

WhatsApp is now a booking channel in India. Kakao, LINE, and WeChat are named as candidates where scale justifies the investment. For agents and intermediaries, that means inventory is moving toward channels they've historically ignored, and commission-relevant search behavior is fragmenting across assistants and messengers rather than consolidating in one interface.

Where is the growth actually happening?

Houldsworth called intra-Asia the fastest-growing travel corridor in the world. Domestic room nights in Asia are growing at a low double-digit rate, against mid-single-digit growth for the company overall. That regional strength has cushioned softness in the Europe–Asia corridor, which she tied to the Middle East conflict.

The measured numbers, in short:

• 64% of surveyed Asia-Pacific travelers use AI for trip planning
• Well under 1% of Booking Holdings' room nights come from direct LLM referrals
• Asia domestic room nights growing low double digits versus mid-single digits companywide
• Alternative accommodation accounts for about 37% of global room nights

What are the risks on the horizon?

Regulation is the main risk Houldsworth flagged for alternative accommodation — the segment's 37% share of global room nights makes it too large to leave exposed to policy shifts. She also framed the connected trip strategy (Booking's push to sell flights, stays, and attractions as one itinerary) as dependent on local payments landing first. Distribution ambitions, in other words, wait on payment rails, not on AI interfaces.

For sellers of travel, the takeaway is blunt. AI is reshaping the top of the funnel, but Booking.com is betting the bottom of the funnel still runs through trust signals, payments, and service — and it's prepared to distribute through WhatsApp and its peers to keep it that way.