AI referral traffic to hotel and short-term rental websites has quadrupled in a year, according to Similarweb data cited in a new Bain & Company analysis — and the advisory firm warns that travel operators that treat AI as a spectator sport risk losing direct bookings, first-party data, and margin.

The stakes are concrete. Bain's Consumer Lab finds that 49% of US consumers who have used AI for researching and booking travel now rely on it most of or all of the time for travel-related tasks. Across the UK, France, and Germany, the figure is 41%. Other travel categories have seen AI's share of web traffic more than double over the same period.

Momentum is building before autonomous, end-to-end booking even matures. Google recently added travel booking functionality to its search AI mode, and Bain expects new tools to keep compounding referral growth from what is still a low base.

"It's clear to us that the entire pre-trip journey is shifting rapidly and that travel operators must make some equally urgent strategic decisions about their response," the analysis states.

Why the direct-vs-indirect split won't save operators

Roughly 70% of travel bookings were made online in 2025, up from 43% in 2011. Yet the split between direct bookings and intermediated bookings has stayed broadly stable across that 15-year span — a fact the wait-and-see camp could read as comfort. Bain calls that wishful thinking.

The stable mix masks massive behavioral shifts beneath the surface, including consumers' embrace of online travel agencies and heavy hotel-brand investment in CRM software to defend direct bookings. "Now isn't the time to ease up on paddling," the analysis warns.

Even in a benign scenario where an operator defends its direct-booking flow as agentic commerce goes mainstream, the economics of the indirect channel may not hold. AI platforms, with their scale and wraparound usefulness, could command fearsome negotiating muscle in travel — comparable, potentially, to Google's leverage in search today. AI agents also enable budget-conscious travelers to hunt bargains systematically, putting direct pressure on pricing.

The harsher scenario: AI tilts the mix toward indirect bookings, cutting operators off from the first-party data that powers AI personalization and diluting earnings.

What should operators actually do?

Bain frames the response around three fronts:

• Content restructured for AI discoverability. Operators must understand how large language models "think" on each platform and which sources they draw on, replacing outdated SEO playbooks. They should still show up as the brand of choice for use cases and segments they have historically won.
• Pricing that AI agents can read. Prices need to be discoverable and understandable by agents, with broader pricing strategy updated to absorb the increased price transparency those agents bring.
• Direct booking that matches the AI experience. That means AI-personalized marketing, stronger loyalty offerings, and customer-facing AI tools — including agents — replicating the convenience of an AI platform. The largest operators may be powerful enough to wall off some inventory from AI search partners or structure joint ventures that mitigate margin dilution.

Marketing budgets, Bain notes, are already gradually shifting toward AI platforms, and operators need new capabilities to follow.

The personalization argument cuts against the industry's traditional caution. Operators of hotels, resorts, theme parks, and cruises succeed by making customers feel welcome and looked-after, and many hesitate to adopt technology that makes the experience less human. Bain argues the opposite move is safer: use AI on the operator's own data stores to build a guest experience memorable enough that the brand comes to mind the next time a traveler asks an AI for a hotel room, resort, park, or cruise.

The questions executives should ask now

Bain poses three framing questions for leadership teams:

• How exposed are we to disruption in customer acquisition, and what is the downside risk of erosion in our direct share over the next 5 to 10 years?
• Should we retain our current place in the travel value chain, or move into adjacencies — and what capabilities would that transition require?
• How should our assets — data, customer relationships, competitive moats — evolve, and to what extent should we partner rather than build?

Smaller brands, Bain acknowledges, are more likely to be optimizing for an agentic AI future than defining it, and pockets of global economic weakness have slowed the industry's post-Covid growth surge. But the analysis is blunt about the alternative: disruption from AI and agentic commerce is coming to every travel provider, whether it helps shape the new booking journey or adapts to a journey shaped by big tech. Operators that wait will live with a booking funnel defined by the market and by technology giants — on their terms, not the operator's.