Google AI Mode

Editorial illustration of travelers flowing through an AI planning funnel that widens at the top but narrows to a trickle at the booking stage, with direct-booking bridges spanning the gap

The AI Booking Gap: Why 56% of Travelers Plan with AI but Book Elsewhere — and What Hotels Must Fix Before 2027

The direct answer: there is a widening chasm between AI-assisted trip planning and AI-completed hotel bookings. Fifty-six percent of travelers have used AI for trip planning (Booking.com survey, 37,000+ respondents, 2026), and 37% use it on travel sites specifically (NYU/BCG, March 2026). But Booking Holdings reported in August 2026 that referrals from ChatGPT, Gemini and other LLMs account for fewer than 1 in every 100 room nights it processes. The gap is not a timing problem that will close on its own. It is a structural breakdown in the path from AI recommendation to completed reservation — and hotels that fix it before 2027 will capture bookings their competitors lose to OTAs every time an assistant says “I found a great hotel for you” but sends the traveler to Booking.com to finish. ...

Deep navy editorial illustration of a glowing map pin hovering over a grid of small hotel icons, streams of light flowing from an abstract AI panel into the pin while most of the light loops back to the panel itself

Google AI Mode for Hotels: Google Keeps 79% of the Clicks — and OTAs Get Almost None

The short answer In Google AI Mode, Google keeps the click — and that is quietly the best news for direct bookings in years. The largest public measurement of hotel results inside Google’s AI Mode (Nicolas Sitter, Google AI Mode Hotel Study 2026, published February 3, 2026) analyzed 4,000 queries across 8 cities — Paris, London, New York, Barcelona, Berlin, Lisbon, Vienna and Bordeaux — capturing 6,130 interactive hotel links and 84,329 source citations. Of every hotel link a traveler can click in an AI Mode answer, 79.1% lands on a Google Business Profile, 16.6% goes directly to the hotel website, and just 3.6% goes to an OTA. Because roughly 97% of those Business Profiles then link onward to the hotel, the study estimates an effective direct-access rate of about 93%. ...

Your Loyalty Program Is Invisible to AI Travel Planners — and That's Where Direct Bookings Die

The short answer Only 38% of AI-assisted hotel bookings complete on the hotel’s own channel. 44% land on an OTA, and 8% complete inside the AI tool itself — according to an NPS Prism and Dynata study of more than 2,300 US travelers who used AI to plan or book a trip (2026). The mechanism is structural, not accidental: AI travel planners build their shortlist of two or three hotels from what they can see from the outside — price and availability — and your loyalty program is precisely the kind of asset that pays off only inside your app, after the choice is already made. Four in ten travelers who booked through AI say it changed which hotel they picked. The scale is no longer hypothetical: Phocuswright (March 2026) puts 56% of active US travelers using AI for planning, booking or in-trip assistance in the previous twelve months, and Adobe Analytics measured AI-referred traffic to US travel sites up 194% year over year in May 2026. The fix is not abandoning loyalty — it is making loyalty legible to machines. With Google’s agentic hotel booking now live in AI Mode (Skift, August 27, 2026) and loyalty pricing entering the answer layer, hotels have a narrow window to do exactly that. ...

AI travel planning interface recommending hotels — editorial tropical-tech illustration

Google AI Mode Can Now Book Hotels: What Independent Hotels Should Do

Google just turned its AI search into a booking engine On August 27, 2026, Google announced that AI Mode in Search can now complete hotel bookings end-to-end. A traveler describes a trip, AI Mode returns a visual list of hotel options with guest reviews and comparison factors, and when one feels right, the traveler taps “Continue on Google”, picks a room, checks the cancellation policy, and pays through Google Pay — without ever leaving the conversation. ...