The short answer
AI travel planners cannot recommend accessible rooms that have no published data behind them — and the industry is finally building the data layer to route around hotels that stay vague. In March 2026, accessible-travel platform Wheel the World closed an $11 million Series A led by Enable Ventures, with participation from Kayyak Ventures, Miles Partnership and — notably — former Expedia CEO Erik Blachford and former Booking.com CEO Gillian Tans (PhocusWire, March 2026). The money funds three things: expanding verified accessibility data with destination marketing organizations, building AI-powered trip planning, and distributing that data to travel platforms and AI tools. “If we consider AI as an engine, the fuel of that engine is data,” says Wheel the World co-founder Alvaro Silberstein.
The demand side already behaves this way. VisitBritain’s 2025 research on travellers with mobility needs finds that detailed information — exact measurements, photos, videos, virtual tours — is not a nice-to-have but essential to booking. Its 2024 survey of 800 travelers with accessibility requirements found that information itself is a barrier to participation: travelers abandon bookings not because properties are inaccessible, but because they cannot verify what “accessible” actually means. When an AI planner faces the same void, it either skips your property or fills the gap with guesswork. If you want to know how your property currently reads inside AI answers, a travel AI audit is the fastest way to find out.
Accessibility was never an infrastructure problem. It was an information problem.
That is Silberstein’s framing, and it is the thesis behind the money. Wheel the World did not raise $11 million to build ramps. It raised it to measure and distribute facts: trained mappers collect more than 200 data points per property — bed height, door width, entrance type, roll-in shower dimensions, restroom details — across 200+ destinations and 8,000 travel services, with 150+ Accessibility Verified partnerships with DMOs across 21 US states (Wheel the World, March 2026). In Oregon alone, the state tourism body has worked with the platform to assess more than 750 hotels, restaurants, parks and transport providers.
Why would the former CEOs of the two largest OTAs on earth invest in a measurement company? Because they know where recommendations are heading. When travelers ask an AI planner “which hotels in Grand Rapids have roll-in showers and step-free restaurant entrances?”, the answer can only be as good as the underlying data. Wheel the World already runs exactly that tool with Experience Grand Rapids and Disability Advocates of Kent County — launched in 2026 and described by the company as the first AI-powered accessibility travel tool of its kind, drawing on verified assessments and covering mobility, vision, hearing and sensory needs. This is the “foundational AI infrastructure for the future of travel” that Enable Ventures’ Regina Kline says she invested in. Structured accessibility data is becoming an input to AI recommendations the way rates and availability became inputs to OTA search two decades ago.
What the research says travelers need — and AI inherits
VisitBritain’s accessibility research program is the most recent rigorous read on what disables a booking. Three findings matter for anyone optimizing for AI recommendations:
- Measurements beat adjectives. The 2025 report on travellers with mobility needs (French, German, Italian and Spanish visitors considering UK trips) concludes that availability of detailed information, measurements and photos/videos/virtual tours is essential. “Accessible” is not a specification; “90cm door width, roll-in shower with fold-down bench, 48cm bed height” is.
- Minimum critical requirements decide shortlists. The 2024 survey of 800 people found accessible parking, accessible toilets and lifts are the baseline features most looked for — the equivalent of a headline price. Properties that don’t state these explicitly don’t enter the shortlist.
- Groups carry multiple requirements at once. Travel parties typically need to meet the needs of several people with several impairments, each with specific requirements. That maps directly onto how people now query AI planners — one long natural-language sentence combining mobility, sensory and dietary constraints — and maps onto nothing in the typical hotel website navigation.
None of this is new to disabled travelers. What changed in 2026 is who reads the void: it is no longer only a frustrated human on your website at midnight, but an AI planner deciding whether your property is a viable answer to cite.
The market math hotels keep getting wrong
The accessible segment is systematically underestimated because most operators count wheelchair users, not travel parties. VisitBritain’s inbound data shows 724,000 visits to the UK in 2023 — 1.9% of all inbound overnight trips — involved someone with a health condition or impairment in the traveling group, generating £624 million in spending (2.0% of all inbound visitor spend). Those visits lasted longer on average than typical inbound trips, with higher total spend per visit. Wheel the World sizes the untapped global accessible travel market at roughly $120 billion (March 2026).
The compounding effect for direct bookings: accessible travelers over-index on properties they can verify before booking, are among the most loyal repeat segments when a trip goes as expected, and — critically for this article — are now among the heaviest users of AI planners, because generic OTA filters have never served their queries well. Every AI recommendation that lands on your verified accessibility page is a booking that costs you zero commission. That is the direct-booking economics our AI visibility program for hotels is built around.
What AI planners do with accessibility queries today
Testing by InsureMyTrip, reported by Forbes in March 2026, compared leading AI models planning a senior-friendly (limited-mobility) Switzerland itinerary: ChatGPT cited fewer sources than other models — “making verification difficult” — and omitted the Swiss Travel Pass, a key tool for that traveler profile, while Gemini surfaced it and structured scenic, lower-mobility alternatives with a price table. The lesson for hotels and operators is not that one model is better. It is that AI answers on accessibility-sensitive queries vary wildly in quality, and the models anchor on whichever sources publish verifiable specifics.
That cuts both ways. We documented in when AI gets your hotel wrong how models invent plausible details for properties with thin data. Accessibility is the highest-stakes version of that failure: an AI that confidently states “roll-in shower available” for a room that has a tub doesn’t just cost you a booking — it can cost a guest their trip and your hotel a refund, a bad review, and in several jurisdictions a legal exposure. Publishing precise facts is hallucination prevention.
The data gap, mapped
| Traveler asks the AI planner | Data the answer requires | Where it lives at most hotels today | What happens when it’s missing |
|---|---|---|---|
| “Hotels in Lisbon with roll-in shower and bed under 50cm” | Shower type and dimensions, bed height, per room type | Nowhere public — front-desk knowledge | Property skipped; a vaguer competitor wins |
| “Step-free route from parking to room, door width for a 66cm chair” | Entrance type, door widths, parking distance | Occasionally a PDF, rarely page-level | Traveler calls three hotels, books the one that answers |
| “Quiet room away from the elevator, low-sensory breakfast time” | Sensory-friendly features, layout notes, policies | Almost never | AI answers generically or invents specifics |
| “Accessible walking tour under 2km on firm surfaces” | Route distance, terrain, rest stops, toilet locations | DMC or guide’s head only | Operator’s itinerary excluded from recommendations |
| “Which restaurants near the hotel have step-free entry?” | Entrance details per venue, verified | Review fragments, inconsistent | Model guesses from reviews — and sometimes wrongly |
The playbook: make your accessibility facts machine-quotable
- Publish an accessibility guide page with real measurements. Bed height, door widths, shower type and dimensions, grab bar positions, entrance type, parking distance, lift dimensions. VisitBritain’s research explicitly recommends an accessibility guide to carry this detail. One page per property, written in plain declarative sentences AI can quote.
- Mark it up. Structured data won’t buy citations by itself — but pages whose markup carries quotable specifications get cited significantly more often, as we showed in hotel schema markup and AI citations. Use
AvailableFeature/ accessibility properties on your LodgingBusiness schema and an FAQ block answering the five queries in the table above. - Add photos of the actual accessible features — the shower, the door, the route from parking. Measurement-plus-photo is the verification pair both disabled travelers and AI systems rely on.
- Get verified. Third-party verification (Wheel the World’s Accessibility Verified is the leading program, with DMO partnerships in 21 states) converts your claims into data that platforms and AI tools can license. Ask your DMO whether a program exists; many fund it.
- Tour operators and DMCs: build accessible itinerary pages with terrain, distance and rest-stop facts. Nobody can recommend the “gentle 2km version” of your tour if it exists only in your best guide’s memory.
- Re-test in the AI tools quarterly. Ask ChatGPT, Gemini and Google AI Mode the exact questions from the table. Log whether your property appears, and whether the details are right. This belongs in the same cadence as your ongoing GEO monitoring — and if you’d rather outsource it, it’s covered in our pricing.
FAQ
Do travelers actually use AI planners for accessible travel? Yes, disproportionately. Accessibility queries are multi-constraint natural-language queries (“roll-in shower, bed under 50cm, quiet floor, 15 minutes from the station”) that legacy OTA filters cannot express and AI planners can. Grand Rapids launched a dedicated AI accessibility trip tool in 2026; Wheel the World raised $11M in March 2026 partly to build more of them.
What accessibility details should a hotel publish at minimum? VisitBritain’s 2024 research points to accessible parking, accessible toilets and lifts as minimum critical requirements, with travelers needing more detail to clarify what each feature comprises. Add per-room measurements: door width, bed height, shower type and dimensions.
Will publishing accessibility data really improve AI recommendations? AI models anchor on verifiable specifics and skip or hallucinate around vague properties. Wheel the World’s entire $11M thesis — endorsed by former Expedia and Booking.com CEOs — is that verified accessibility data is the fuel AI travel planning runs on.
How big is the accessible travel segment for a single hotel? VisitBritain counts 724,000 inbound UK trips in 2023 (1.9% of visits, £624 million) involving a traveler with a health condition or impairment, with longer-than-average stays and higher spend per visit. Wheel the World estimates the untapped global market at $120 billion.
Where does this fit against OTA-driven demand? Accessible travelers book where they can verify. An AI answer that cites your accessibility guide converts to direct bookings at zero commission — the same direct-versus-OTA economics as every other GEO play, with a more loyal repeat guest on the other end.
Accessibility is the clearest case in travel of a segment whose bottleneck was never demand or inventory — it was information. AI planners just made information the whole game. The properties that publish measurements will be recommended; the ones publishing the word “accessible” will be skipped. Which side of that line you sit on is, for once, entirely your choice.