Visuals Are Data Capture: Why the 6% AI Booking Gap Rewards Properties That Capture Their Space

By Jascha Kaykas-Wolff, CEO, Visiting Media (with A.J. Hamil, VP of Product Studio)

AI drives 37 percent of travel discovery and research touchpoints yet only 6 percent of bookings. Because 94 percent of bookings still happen outside the AI chat after the buyer checks the property for themselves, investing in visuals of your space is about structured data capture, not flat photography.

A recent Skift Research study of the travel booking journey found that AI accounts for 21 percent of touchpoints at the inspiration stage and 37 percent across inspiration and research, yet drops to just 6 percent at the moment of booking. Across major global brands and OTAs, direct AI-assisted checkout still represents less than 1 percent to roughly 1.6 percent of total room nights.

Some hotel operators look at that 6 percent booking figure and conclude they have another two years before AI search matters to their commercial results. I think that reading misses what the data actually shows.

If 37 percent of discovery and research touchpoints already run through AI systems and only 6 percent end in an in-chat transaction, then 94 percent of bookings still happen outside the AI chat, after the buyer checks the property for themselves. In research Booking.com shared at HumanX Europe, 89 percent of consumers said they want to use AI to research travel, while only 6 percent said they trust it to make the final decision. Hospitality Net’s analysis of AI shortlist behavior shows the same pattern: only 5 percent of travelers book blindly from an AI recommendation, and roughly half click straight through to the hotel’s own website to verify the property for themselves.

Two gates every property now has to clear

That 31-point gap between AI discovery (37 percent) and AI checkout (6 percent) splits the hospitality buying journey into two distinct gates. First, your property has to be understood by the machine so it makes the three-to-five hotel shortlist when a traveler or event planner asks a multi-constraint question. Second, when that skeptical buyer clicks out of the chat interface to inspect your property with their own eyes, your visual experience has to close the trust gap.

Even after a guest books, the visual verification test keeps running. A joint October 2026 survey by Skift Research and McKinsey of 1,015 US travelers found that AI touchpoints rebound to 29 percent post-booking, 41 percent of travelers keep monitoring prices after reserving a room, 29 percent keep shopping for a better option, and 51 percent have switched hotels after seeing a preferable option on social media.

Why investing in visuals is about data capture, not photography alone

For twenty years, hotels treated visual production as a cosmetic marketing expense. You hired a photographer to shoot flat 2D pictures for a digital brochure, uploaded the JPEGs to a gallery page and called the job done.

In an era where software builds the shortlist and humans verify the physical reality, every property has to invest in strong visuals of its space, and that investment is fundamentally about structured data capture rather than photography alone. A flat JPEG is a single frame for a human eye. A verified 3D and 360 spatial capture records the underlying geometry of the property: room dimensions, ceiling heights, sightlines, terrace adjacencies, banquet capacities and spatial relationships across the grounds.

“The 6 percent AI booking stat isn’t a reason for hotels to wait. It’s proof that 94 percent of bookings still happen outside the AI chat, where buyers verify the space with their own eyes. Every property has to invest in strong visuals of its space, and today that investment is about structured spatial data capture, not flat photography alone.”

Jascha Kaykas-Wolff, CEO, Visiting Media

When we audited 1,500+ hospitality websites with our Visual Quotient (VQ) benchmark, we saw what happens when properties stop at flat photography. Across our 1,203 audited customer properties, the average VQ score is 34.5 out of 100, while non-customer properties average 27.9 out of 100. More than four out of five websites (83.9 percent) publish zero machine-readable spatial schema for answer engines, and 97.9 percent lack interactive floor plans on the public web.

Capturing physical reality once to serve both machines and people

Inside our Product Studio, A.J. Hamil, our VP of Product Studio, and the team build around a simple rule: capture the physical property with precision once, and let that spatial record power every surface where a buyer discovers or evaluates the hotel.

“When we capture a property in 3D and 360, we’re not just producing pictures for a gallery. We’re capturing a spatial dataset: the room’s geometry, scale and layout, and how each space connects to the next. Capture it accurately once, and that same record can feed the structured data an AI search tool needs to recommend the property, and the interactive walkthrough a planner needs to trust the booking.”

A.J. Hamil, VP of Product Studio, Visiting Media

When a property treats spatial capture as foundational data rather than a one-off photo shoot, one verified 3D and 360 record does the work across the entire buying journey. It gives answer engines the structured spatial facts they need to recommend a room or ballroom accurately, and it gives leisure guests and event planners the interactive walkthrough and floor-plan truth they need to book with confidence.

The 6 percent AI booking figure isn’t a grace period. The shortlist is already forming inside AI systems, while the final decision still belongs to a human buyer checking the space for themselves. Hotels need to think about whether their most valuable physical inventory is actually visible and understandable to the software narrowing the options and convincing to the people making the choice.