A boutique hotel in Dallas does not lose a direct booking to a better OTA listing. It loses it to a niche, hyper-local task no generic page answers. The guest asks an AI assistant for a quiet room near the Design District with late checkout and a walkable coffee shop. If the answer engine cannot find property-level proof, it falls back on the sources it already trusts. Those sources are usually OTAs, guides, and listicles.
Hotels need proof before they need more traffic. Every room night shifted from an OTA to direct saves the spread between 15 and 20 percent commission and a 5 percent planning assumption for the cost of taking a direct booking. At 18 percent commission and a 5 percent direct-cost assumption, a $650 booking keeps $84.50 more. A $180 booking keeps $23.40 more.
The AEO return that survives a CFO review
Start with the spread. Lighthouse reports OTA commissions typically run 15 to 20 percent per booking. Booking.com says the exact rate depends on country, property type, and agreement, so treat it as a planning range and replace it with contract data. Direct bookings are not free. Card processing, booking engine fees, and marketing allocation still cost money. Stripe publishes card rates from 1.5 percent plus a fixed fee for European cards to 3.15 percent plus a fixed fee for international cards. A 5 percent all-in direct-cost assumption is reasonable if labeled as an assumption.
| Metric | 15% OTA | 18% OTA | 20% OTA |
|---|---|---|---|
| OTA commission per $100 of room revenue | $15 | $18 | $20 |
| Direct cost per $100 at a 5% planning assumption | $5 | $5 | $5 |
| Net retained per $100 shifted | $10 | $13 | $15 |
| Net retained per shifted $180 booking | $18 | $23.40 | $27 |
| Net retained per shifted $650 booking | $65 | $84.50 | $97.50 |
A hotel keeps 10 to 15 cents of every room-revenue dollar moved from an OTA to direct.
Apply that spread to two illustrative properties. A 30-room boutique at $180 ADR and 72 percent occupancy grosses $1.42 million. A 10-point channel shift retains $18,449 at 18 percent commission and 5 percent direct cost, leaving $3,449 against a $15,000 program. A 50-room resort at $650 and 68 percent occupancy grosses $8.07 million. The same shift retains $104,865, leaving $64,865 against $40,000.
| Property | Gross room revenue | 10-point OTA-to-direct shift | Net retained at 18% OTA and 5% direct | Program cost | Break-even shifted share |
|---|---|---|---|---|---|
| 30-room boutique, $180 ADR, 72% occupancy | $1,419,120 | $141,912 | $18,449 | $15,000 | 8.1% |
| 50-room luxury resort, $650 ADR, 68% occupancy | $8,066,500 | $806,650 | $104,865 | $40,000 | 3.8% |
ADR, OTA dependence, and the commission spread drive the return more than AI traffic does. At 18 percent commission and 5 percent direct cost, every dollar of program cost needs $7.69 of shifted room revenue to break even. That ratio does not change with ADR. ADR changes how many bookings it takes.
The estimate leaves out elasticity, cancellations, upsells, taxes, and OTA guests who would never book direct. If a hotel cannot shift the break-even share, AEO is a brand project, not a direct-booking investment.
What the 2026 data shows about AI hotel discovery
AI referrals are still small. Lighthouse found AI referral traffic was just under 1 percent of hotel website visits. Similarweb data reported by Hospitality Net showed AI referral share for hotel and short-term rental sites quadrupled between July 2025 and July 2026, then stayed under one visit in a hundred. Booking Holdings said AI traffic was significantly below 1 percent of its room nights.
The audience is bigger than the referral share suggests. Phocuswright found 56 percent of US leisure travelers used AI for a trip in the previous twelve months, up from 43 percent in the second half of 2025 and 33 percent in the first half. Bain found 49 percent of US travelers who have used AI for travel now rely on it most or all of the time. Adobe data reported by Hospitality Net showed AI-referred visitors to US travel sites converted roughly as well as other visitors by July 2026, closing a 47 percent gap. Traffic is tiny, but the intent behind it is strong.
The widely quoted 11.4 percent AI conversion against 5.3 percent organic is a Similarweb aggregate across the open web, not a hotel number. Travelers convert below that average, and they typically research a stay over weeks.
Stiplo analyzed 659 Barcelona accommodations and 27,360 AI answers. Independent hotels were 8 times more likely than chains to be completely unknown to AI. About 1 in 5 was never recommended, and about 20 percent captured more than 80 percent of mentions.
Look at AI's sources. About 68 percent of what AI said came from guides and listicles. Only about 7 percent came from hotel websites, and only about 7 percent of recommendations linked to the hotel's own site. Most pointed to OTAs, with Booking.com named most often, then TripAdvisor and listicles. About a third came from model memory. The study found no measurable independent effect from llms.txt, FAQ markup, or schema packages. Google review volume was the strongest driver, followed by backlinks and rankings, guide features, and clear content.
BrightEdge found Expedia in 31.3 percent of Google AI Overviews answers and 26.6 percent of ChatGPT answers for US holiday travel, with Booking.com at 13.6 and 12.9 percent. The answer layer is OTA-shaped because OTAs already publish inventory, price, and availability.
Independents and luxury properties have an opening. AI does not need another city description. It needs evidence an OTA template cannot supply. The Barcelona study also found that once an AI knew an independent hotel, it recommended it about as often as a chain. The problem is being unknown.
The commission math is simple, the attribution is not
Track three numbers. Shifted room revenue is revenue that would have gone to an OTA but now books direct. Incremental revenue is demand that would not have existed without the work. Program cost is content, creator activation, tooling, and staff time. Direct cost sits inside the spread.
If AEO only captures guests who were already going to book direct, the return is zero. If it captures guests who saw the property in an AI answer, then checked the OTA listing and booked there, the return is negative. Last-click attribution undercounts a proof layer that influences the shortlist before the booking engine sees a session.
Use three signals together. Pull AI referrals in analytics for chatgpt.com, perplexity.ai, gemini.google.com, and copilot.microsoft.com, then compare their conversion rate with organic search. Tag direct bookings with a rate code or booking engine source and count shifted room nights, not clicks. Add an AI option to the "how did you hear about us" field and watch branded search in the property's feeder markets. Together they show whether commission avoidance is real.
Why OTA-shaped answers leave a niche opening
OTA pages are strong on inventory, price, and availability. They are weak on the details that decide a stay. A guest asking for a quiet room, a 2 PM checkout, a walkable breakfast spot, or a suite for two kids wants property-level proof. Guides and listicles win those answers because they describe experience.
A hotel can build that proof with local creators who already publish the exact content the task needs. A million-follower beach reel is not evidence that your Dallas property has a quiet courtyard and good espresso.
Start with a brief that names the property, neighborhood, guest task, and season. Break it into three to six niche topics, such as a quiet business stay, a wellness weekend, or a family suite near transit. Search actual creator content instead of profile filters. Disqualify creators whose latest content does not match the task. Validate location from three independent sources, the creator's declared location, public information, and recent published content. Build a fit profile with reputation, competitor work, and a verified contact route. Let the hotel approve the validated creator before outreach. Then generate one-to-one TikTok and email outreach from the task context and the creator's content theme. Charge only after the hotel approves a validated creator.
That workflow is what Lobby builds. It is niche and hyper-local creator activation infrastructure. It is not an AEO audit, a booking engine, or a guest CRM. Lobby does not replace your property management system, booking engine, or guest CRM, and it does not guarantee citations, bookings, or ROI.
AI systems pick up creator proof from the third-party surfaces the study found dominant. A local creator video is evidence that a specific room, street, service, or seasonal detail exists, in a form an AI system can parse and a guest can trust. Pair that with review volume, guide placements, and a fast booking path, and the property has something an OTA page cannot copy. The AEO for luxury hotels guide covers the search side. The modern hotel marketing tech stack guide covers the technical layer, and the hyperlocal geo-search guide explains why city filters miss the neighborhood creator.
If the booking engine is slow, the rate is uncompetitive, or the front desk cannot honor the late checkout the creator filmed, proof only feeds a leaky funnel. AEO supports demand and evidence work, not operations.
A 90-day measurement plan that a hotel can defend
Days 1 to 30 build a baseline. Pull twelve months of AI referrals and compare conversion with organic search. Audit branded search in top feeder markets. Add an AI option to the booking form's attribution question. Pull OTA mix and commission by room type. Calculate the break-even share of room revenue. If it exceeds what the hotel can shift, stop.
Days 31 to 60 build task proof. Build three to six topic pages, one per guest task, not one per amenity. Commission creators whose recent content matches. Publish with property facts, rates, and a clear direct booking path. Track citations for 20 high-intent queries across ChatGPT, Gemini, and Perplexity.
Days 61 to 90 deliver the verdict. Count shifted room nights, assisted direct revenue, and commission avoided. Keep topics that produce direct demand. Kill the ones that only produce views. Do not judge AEO on AI referral sessions alone. That number will look tiny even when the math works.
To test the creator-proof layer without a retainer, start with a property brief in Lobby. Lobby decomposes the brief into niche topics, finds creators who already publish that content, disqualifies creators whose latest content does not match the task, validates location three ways, and lets your team approve before anyone is contacted. You pay per activation, and only after you approve a validated creator. Start free activation. For a broader set of hospitality activation patterns, explore 700 real activation strategies.
Frequently asked questions
How do you calculate AEO ROI for a hotel
Take the room revenue you expect to shift from OTA to direct, multiply it by the difference between OTA commission and direct booking cost, then subtract program cost. At 18 percent commission and 5 percent direct cost, every $100 shifted keeps $13. Add incremental direct bookings, and do not count bookings that would have come direct anyway.
Is AI traffic big enough to justify AEO in 2026
Not on its own. AI referrals are under 1 percent of hotel website visits. The justification is commission avoided on shifted demand, plus early citation presence while the answer layer is still forming. If your ADR is low and OTA dependence is low, the break-even share may be too high to justify a large program.
Do schema markup and llms.txt improve hotel visibility in AI search
Not by themselves. The Barcelona study found no measurable independent effect from llms.txt files, FAQ markup, or schema packages. Those formats help machines parse facts, but they do not create proof. The stronger drivers were Google review volume, backlinks and rankings, guide features, and clear content that answers real guest questions.
How long does AEO take to affect direct bookings
There is no honest fixed timeline. Travel booking windows run for weeks, AI memory is built over months, and citations shift as models change. Use 90 days to measure leading indicators such as citations, AI referrals, branded search, and assisted direct revenue. Give commission avoidance six to twelve months before a fair read.
Can Lobby replace a booking engine or an AEO agency
No. Lobby is niche and hyper-local creator activation infrastructure. It does not replace your property management system, booking engine, guest CRM, or AEO audit, and it does not guarantee citations or revenue. It decomposes a property brief into niche topics, finds creators who already publish that content, disqualifies creators whose latest content does not match, validates location three ways, and helps the hotel approve and contact the right creator.
Tired of static influencer databases?
Lobby replaces dead directories with live TikTok creator search and direct outreach. Zero manual vetting, verified contacts, and live engagement metrics.