If you want to get your hotel recommended by Claude, a useful starting point is understanding when it appears in travelers’ searches and how accurately it is described. Americas Great Resorts compared 80 Claude hotel answers across six destinations to examine which properties made the shortlist, how recommendations changed when questions were repeated, and which important property details survived.
Claude changed at least one recommended hotel in 25 of 30 repeated hotel discovery questions. Most of those changes were partial: 19 questions kept two of their three hotels. One New York couples’ getaway question returned three entirely different properties.
These answers were collected on October 6, 2026. They reveal two separate things worth tracking: whether Claude recommends a property and whether its description preserves the facts a traveler needs.
For broader steps to improve your property’s AI visibility, see our guide to getting your hotel recommended by AI.
Hotel Wailea’s descriptions consistently retained important age and location distinctions. In another case, a published January closure notice affected the first run but disappeared from the second, even though the requested stay dates had not changed. The hotel gained a place on the shortlist while the answer lost a potentially important qualification.
Here, “finds” means the hotels and supporting information visible in Claude’s answers. The research examines those outputs, without claiming access to Claude’s internal selection process.
What we asked Claude about hotels
We asked the same 40 questions in two fresh sessions: 30 discovery questions and 10 named-hotel questions per run, producing 80 answers. Both question types requested supporting sources and links, with no prices or room-availability claims.
Discovery covered six markets: New York City, the Miami area, the Los Angeles area, Chicago, Napa Valley, and Maui. Each had five traveler purposes: overall luxury leisure, a romantic couples’ getaway, a family holiday, wellness, and distinctive local character.
Each discovery question requested three hotels for January 15–17, 2027. The Miami area included Miami Beach and Surfside; the Los Angeles area included Beverly Hills.
The named-hotel questions covered operating status, location, pool, spa, and family suitability. They examined descriptions when the hotel was supplied by name and were excluded from discovery appearance counts.
How often Claude’s hotel recommendations changed
Across the 60 discovery answers, Claude supplied 180 recommendation appearances involving 52 distinct hotels. We matched variations in property names before comparing the two runs and treated a reordered list of the same three hotels as unchanged.
| Market | Discovery answers across both runs | Distinct hotels recommended | Questions with the same three hotels in both runs |
|---|---|---|---|
| New York City | 10 | 12 | 0 of 5 |
| Miami area | 10 | 8 | 2 of 5 |
| Los Angeles area | 10 | 9 | 1 of 5 |
| Chicago | 10 | 9 | 0 of 5 |
| Napa Valley | 10 | 8 | 1 of 5 |
| Maui | 10 | 6 | 1 of 5 |
| Total | 60 | 52 | 5 of 30 |
Only five questions returned the same three properties twice. Nineteen retained two properties, five retained one, and one retained none.
The distinction matters. Saying that 25 shortlists changed is accurate. Saying that Claude started from scratch 25 times would misrepresent the results. In most comparisons, there was a recurring pair alongside a different third choice.
New York’s couples shortlist changed completely
For the New York romantic getaway question, the first run recommended The Lowell, Aman New York, and The Whitby Hotel. The second recommended Baccarat Hotel New York, The Carlyle, and The Greenwich Hotel.
The destination, occasion, dates, and request for three hotels were identical. None of the first three names returned in the second answer.
A hotel team looking only at the first answer could have concluded that the property had secured a place in this particular conversation. A team looking only at the second could have reached the opposite conclusion. The pair shows why a single answer is an incomplete view of visibility.
Miami showed a more persistent group of hotels
The Miami overall-luxury question returned The Setai, Four Seasons Hotel at The Surf Club, and Faena in both runs. The wellness question also kept the same trio across runs: Carillon Miami Wellness Resort, Faena, and The Setai.
Across all ten Miami discovery answers, The Setai appeared nine times, The Surf Club seven times, and Faena six times. These are appearances within this specific question set, not a market-wide ranking or an estimate of traveler demand.
Together, New York and Miami show why the market and question belong beside the result. Repetition can reveal a persistent group of hotels, substantial changes, or both. A single site-wide “AI visibility” label would conceal those differences.
The reason for the trip changed the relevant shortlist
The Maui couples question retained Hotel Wailea, Four Seasons Resort Maui at Wailea, and Hāna-Maui Resort in both runs. The family question retained Grand Wailea and Four Seasons, while its third recommendation changed from The Ritz-Carlton Maui, Kapalua to Fairmont Kea Lani.
Hotel Wailea appeared in both couples answers and neither family answer. Four Seasons appeared in both sets of questions in both runs. The answers therefore captured two different patterns: a property associated with a narrower occasion and a property presented as suitable for more than one type of trip.
The reasons also differed. In the couples answers, Claude emphasized Hotel Wailea’s adults-only overnight policy and setting. In the Four Seasons descriptions, it distinguished adult-oriented facilities from the wider family resort experience. The family answers discussed children’s activities, pools, and accommodation features.
For hotel teams, the practical question is specific: does the property appear for the occasions it actually serves, and does the answer explain the right reasons? An absence from a family shortlist can be appropriate for an adults-only retreat. An absence from both couples shortlists would raise a different question for that same hotel.
How accurately Claude preserved important hotel details
The named-hotel questions covered ten properties in each run. We checked selected consequential details against the hotels’ own published information, looking for accurate descriptions as well as omissions. Three examples show why the wording around an amenity or policy deserves attention alongside the hotel name.
Hotel Wailea: location and age distinctions survived
Both named-hotel answers described Hotel Wailea as a hillside property rather than a beachfront hotel. They explained that Wailea Beach is approximately a five-minute drive away and that guest transportation is available. Both also identified the minimum overnight guest age as 18.
Those details agree with the hotel’s official FAQ. They preserve a meaningful distinction for a traveler choosing between a hillside retreat and direct beach access.
The first answer also retained the exception allowing younger guests at the restaurants. Both noted that weddings and private events may include guests of all ages. The adults-only overnight policy was therefore not expanded into a blanket claim that children could never enter the property.
The spa descriptions were more cautious. Both reported in-room spa services without establishing a dedicated spa facility. The hotel’s experiences page confirms an in-room spa menu; that alone does not establish a separate spa building.
This was a useful outcome: accurate central facts accompanied by uncertainty where the available information did not settle a different question.
The Peninsula Chicago: a small qualifier changed the meaning
Both runs correctly described The Peninsula Chicago’s pool as indoors. Its official page also identifies an indoor hot tub and an outdoor Sun Terrace, keeping the swimming facility separate from the outdoor space.
The family schedule received slightly different treatment. The first run called the published times “family hours” and said access outside them was not established. The second explicitly preserved the hotel’s wording that these were recommended family hours, rather than a stated ban at other times.
The official page supports the second formulation. It recommends 9 a.m.–1 p.m. and 3–7 p.m. for families.
The first answer did not explicitly prohibit children outside those windows. The difference is one of precision: keeping “recommended” helps prevent a reader from interpreting guidance as a restriction. For a family organizing a short stay, that qualifier can matter more than another sentence praising the pool.
Auberge du Soleil: a seasonal warning disappeared
Auberge du Soleil produced the most consequential discrepancy among these examples.
In the first run, Claude excluded the hotel from its Napa overall-luxury and couples recommendations because its official policy page listed a January seasonal closure. The named-hotel answer preserved the uncertainty: the page did not give exact 2027 closure dates, so it did not establish whether January 15–17 was affected.
In the second run, Claude recommended Auberge du Soleil for both questions. It cited the same policy page for other facts, including pool access and the hotel’s adult-oriented positioning, but did not carry forward the January warning. Its named-hotel answer said no closure notice had been found.
Our check of the official policy page found January still listed under seasonal closures. Exact 2027 dates remained unspecified.
The verified issue is the omitted notice. We have not confirmed that the hotel will be closed on the requested dates. The second run nevertheless gave a traveler less reason to investigate a potentially important limitation than the first run did.
This example also connects the description to the shortlist. The first answer explicitly used the closure notice to exclude the property. The second included it without that qualification. A hotel can gain a recommendation appearance while the answer becomes less complete.
What the supporting links did and did not establish
Every one of the 80 answers contained supporting links, as requested. The reports linked to hotel and brand websites, travel publications and guides, destination information, and booking listings. That gave us specific pages to check against the statements beside them.
Official information supported useful details in the Hotel Wailea and Peninsula examples. It also exposed the missed qualification at Auberge du Soleil. A link to the right property page was therefore useful evidence for review, but did not guarantee that the answer retained every relevant fact on that page.
The two runs also differed in what they said they could establish. In the first Peninsula Chicago answer, Claude reported that it could not read an official family-program page and relied partly on other websites. In the second, it linked to an official family-amenities page and listed the services described there.
The reports describe different stated access outcomes. Before commissioning a technical fix, have the web team check whether the reported page-access failure recurs and what is causing it.
Recommendation, description, and link destination remain separate observations. A property can be recommended while a supporting link points to a travel guide. A property website can be linked while an important qualification is omitted. Neither observation measures a visit or booking. AGR’s separate analysis of who receives the link in AI hotel recommendations explores that distinction further.
How to check whether Claude recommends your hotel
Start with discovery questions that describe your destination and the trips your guests plan, without naming your hotel. Then ask about your property directly. Repeat the questions in a fresh session and compare both the recommended hotels and the details attached to them.
The changing New York shortlist and the omitted Napa closure warning call for different responses. One requires a broader view of where the hotel appears; the other requires checking whether the answer preserves a material fact. The actions below follow those distinctions.
| Finding in these Claude answers | Action for the hotel team | Who should verify it |
|---|---|---|
| 25 of 30 repeated shortlists changed at least one hotel | Keep the question fixed and compare repeated answers before judging visibility | Marketing |
| Couples and family questions produced different property combinations | Check the occasions the hotel genuinely serves and the reasons given for inclusion | Marketing and reservations |
| A January closure warning disappeared while other facts from the policy page remained | Publish precise affected dates and reconcile seasonal notices with amenity and booking information | Operations and the website owner |
| Recommended family pool hours needed a qualifier | State whether a schedule is advisory or mandatory, including exceptions | Operations |
| In-room spa services did not establish a dedicated spa facility | Describe the service and its location explicitly | Spa or guest-services team |
Start with the discrepancy that could most affect a booking decision. A seasonal closure, an age restriction, or a location misunderstanding deserves attention before a flattering adjective or a change in list order.
For each issue, keep the original question, the answer, and the relevant property page together. Have the operational owner confirm the fact before marketing changes the wording. If the hotel’s own page is ambiguous, correct it there; if another publisher carries an incorrect fact, request a correction with supporting information.
Then repeat the relevant question and compare the actual wording. A corrected page and a changed Claude answer are separate things to verify. This research did not test an intervention or establish how quickly a website correction would appear in later answers.
What these 80 answers tell us about Claude hotel visibility
This collection contains two runs of the same questions on one date, with each run answering the full set in one session. It shows what Claude returned for these six markets and five traveler purposes. It does not establish every hotel’s visibility, performance across other AI systems, or a trend over time. Supporting links were requested, so their presence should not be treated as a measure of how often Claude supplies them unprompted.
Within that scope, the findings are concrete. Most repeated shortlists changed at least one property, but most also kept two. The property checks showed why descriptions also deserve attention: a correct hotel name can be accompanied by precise details, incomplete information, or an omitted qualification.
For a hotel team, visibility has two parts worth tracking together: inclusion in the relevant traveler conversation and accuracy in the explanation that follows. Recording both provides a more useful assessment than celebrating a mention or reacting to a single absence. Hotel teams can use those same two questions when reviewing their presence on other AI platforms.
Independent luxury hotel and resort teams can request a free AI Visibility Audit to see how ChatGPT, Gemini, and Google AI Mode describe their property. Findings arrive within five business days and cover observed AI visibility and errors in the property’s public record.
