Which Hotels Do AI Engines Recommend? A Study of 10 Destinations

Americas Great Resorts Research | September 27, 2026

Americas Great Resorts analyzed 1,177 AI-generated hotel recommendation answers across 10 destinations, covering 100 distinct travel questions asked three times each on ChatGPT, Google AI Mode, Gemini and Copilot. The September 27, 2026 study identified 6,110 featured hotel appearances, measuring which properties surfaced most often, how recommendations differed between engines, and how the traveler’s question changed the results.

Findings at a glance

  • Engine choice mattered: Nine of ten destinations had different leading hotels across the four engines. Honolulu was the exception, with Halekulani leading all four.
  • Travel intent changed the leaders: Orlando’s most frequently featured hotels differed for family vacations, boutique stays and Walt Disney World visits.
  • Repeated questions produced variation: Hotel-list membership changed in 363 of the 379 question-and-engine groups with three assessed answers.
  • Some hotels appeared consistently: The Peninsula Chicago featured in 96 of 118 Chicago answers; Halekulani appeared in 91 of 118 Honolulu answers.

These findings measure visibility within a fixed set of travel questions, rather than hotel quality, booking performance or the likelihood that every traveler will receive the same recommendations. AGR’s hotel AI visibility guide explains the broader distinction between being recognized by name and being included in a traveler’s hotel choices.

This report establishes a dated baseline for this ten-destination question set. It complements the earlier AGR Luxury Hotel AI Visibility Index, which examined six U.S. markets across three engines. The studies use different designs and counting methods, so their totals should not be treated as a direct measure of change over time.

The hotels that appeared most often in each destination

The study covers ten destination areas and ten questions per destination. Eight question types are shared across the destinations: general hotel discovery, luxury, couples, families, boutique hotels, spas, dining and service. Two additional questions address local travel needs.

Each question was repeated three times on each of the four engines. The table below combines all ten questions and all four engines within each destination. A hotel receives at most one appearance per answer, even if it is named several times. A clearly identified hotel's spa recommendation contributes to that hotel's visibility on the spa question.

The three most frequently featured hotels in each destination are listed in descending order. Each hotel has its own row, with appearances shown separately from its name.

DestinationHotelAppearances
New York CityMandarin Oriental, New York50/120
New York CityThe Peninsula New York36/120
New York CityRitz-Carlton New York, Central Park35/120
Miami areaThe Setai, Miami Beach51/117
Miami areaFaena Miami Beach49/117
Miami areaFour Seasons Hotel Miami45/117
Las VegasWynn Las Vegas67/116
Las VegasBellagio Hotel & Casino66/116
Las VegasWaldorf Astoria Las Vegas59/116
ChicagoThe Peninsula Chicago96/118
ChicagoThe Langham, Chicago67/118
ChicagoFour Seasons Hotel Chicago66/118
Los Angeles areaHotel Bel-Air54/116
Los Angeles areaThe Beverly Hills Hotel52/116
Los Angeles areaFour Seasons Los Angeles at Beverly Hills49/116
Orlando areaFour Seasons Orlando77/120
Orlando areaRitz-Carlton Orlando, Grande Lakes75/120
Orlando areaWaldorf Astoria Orlando60/120
Cancún / Costa MujeresExcellence Playa Mujeres64/116
Cancún / Costa MujeresAtelier Playa Mujeres54/116
Cancún / Costa MujeresLe Blanc Spa Resort Cancun50/116
Punta CanaTortuga Bay56/117
Punta CanaEden Roc Cap Cana53/117
Punta CanaSecrets Cap Cana Resort & Spa49/117
Los CabosWaldorf Astoria Los Cabos Pedregal72/119
Los CabosOne&Only Palmilla64/119
Los CabosMontage Los Cabos61/119
Honolulu / WaikīkīHalekulani91/118
Honolulu / WaikīkīThe Kahala73/118
Honolulu / WaikīkīThe Royal Hawaiian73/118

Appearances are shown out of the assessed answers for that destination: 50/120 means the hotel featured in 50 of 120 assessed answers. The Kahala and The Royal Hawaiian tie on 73 appearances. Small differences, including Wynn's 67 versus Bellagio's 66, should not be interpreted as durable advantages.

The table shows two useful patterns. Some markets have a pronounced observed leader: The Peninsula's 96 Chicago appearances compare with 67 for The Langham. Other markets have closely grouped leaders, including Las Vegas and Orlando.

The results also show why a single overall hotel ranking would be misleading. Each destination has its own questions and competitive context. A Chicago appearance count and a Miami appearance count do not establish which hotel is better, or which has more visibility across the internet.

The study's wording is deliberately oriented toward upscale travel. Luxury, exceptional service, dining, spas and romantic getaways feature prominently. These results therefore describe that question set; they do not measure the full economy, motel or budget hotel market.

Different AI engines produce different hotel leaders

Pooling engines is useful for seeing repeated visibility, but it conceals substantial differences. The table below identifies the hotel with the most appearances on each engine in each destination. Each tied hotel has its own row with the same appearance count.

DestinationAI engineLeading hotelAppearances
New York CityChatGPTMandarin Oriental, New York18/30
New York CityGoogle AI ModeThe Greenwich Hotel19/30
New York CityGeminiMandarin Oriental, New York14/30
New York CityCopilotRitz-Carlton New York, Central Park15/30
Miami areaChatGPTThe Setai, Miami Beach20/29
Miami areaChatGPTFaena Miami Beach20/29
Miami areaGoogle AI ModeFour Seasons Hotel at The Surf Club14/28
Miami areaGeminiThe Setai, Miami Beach15/30
Miami areaCopilotRitz-Carlton Key Biscayne, Miami21/30
Las VegasChatGPTFour Seasons Hotel Las Vegas27/28
Las VegasGoogle AI ModeWaldorf Astoria Las Vegas22/28
Las VegasGeminiBellagio Hotel & Casino21/30
Las VegasCopilotBellagio Hotel & Casino24/30
ChicagoChatGPTThe Langham, Chicago28/30
ChicagoGoogle AI ModeThe Peninsula Chicago21/28
ChicagoGeminiThe Peninsula Chicago25/30
ChicagoCopilotThe Peninsula Chicago25/30
Los Angeles areaChatGPTHotel Bel-Air21/28
Los Angeles areaGoogle AI ModeThe Beverly Hills Hotel14/29
Los Angeles areaGeminiThe London West Hollywood11/29
Los Angeles areaCopilotFour Seasons Los Angeles at Beverly Hills20/30
Orlando areaChatGPTFour Seasons Orlando27/30
Orlando areaGoogle AI ModeRitz-Carlton Orlando, Grande Lakes18/30
Orlando areaGeminiJW Marriott Orlando, Grande Lakes19/30
Orlando areaCopilotRitz-Carlton Orlando, Grande Lakes20/30
Cancún / Costa MujeresChatGPTAtelier Playa Mujeres20/29
Cancún / Costa MujeresGoogle AI ModeAtelier Playa Mujeres20/28
Cancún / Costa MujeresGeminiExcellence Playa Mujeres21/29
Cancún / Costa MujeresCopilotLe Blanc Spa Resort Cancun18/30
Punta CanaChatGPTTortuga Bay17/28
Punta CanaChatGPTSecrets Cap Cana Resort & Spa17/28
Punta CanaGoogle AI ModeTortuga Bay17/29
Punta CanaGeminiEden Roc Cap Cana14/30
Punta CanaCopilotEden Roc Cap Cana18/30
Los CabosChatGPTMontage Los Cabos28/29
Los CabosGoogle AI ModeWaldorf Astoria Los Cabos Pedregal20/30
Los CabosGeminiWaldorf Astoria Los Cabos Pedregal15/30
Los CabosGeminiNobu Hotel Los Cabos15/30
Los CabosCopilotOne&Only Palmilla18/30
Honolulu / WaikīkīChatGPTHalekulani27/30
Honolulu / WaikīkīGoogle AI ModeHalekulani26/30
Honolulu / WaikīkīGeminiHalekulani15/28
Honolulu / WaikīkīCopilotHalekulani23/30

The Appearances column shows the hotel's appearances out of the assessed answers for that engine and destination. It is an observed answer frequency, not market share or a predicted probability for another user.

New York illustrates the distinction. Mandarin Oriental led on ChatGPT and Gemini. The Greenwich Hotel led on Google AI Mode, while The Ritz-Carlton New York, Central Park led on Copilot. A hotel monitoring only one of those engines would see an incomplete picture of this study's results.

Chicago offers a different pattern. The Peninsula led the combined count and three engines, but The Langham led ChatGPT. Honolulu showed the strongest agreement at the top: Halekulani led all four engines, although its appearance counts varied between them.

Agreement on a leader does not mean agreement on the rest of the hotels. Nor does a different leader establish that one engine is more accurate. This study measures the hotels presented, rather than scoring the engines' travel advice.

How travel intent changes hotel recommendations: an Orlando example

The same destination can produce a different leader when the travel requirement changes. Orlando provides an illustrative example, with twelve assessed answers for each of its ten questions.

Orlando travel question / intentMost frequently featured hotelAppearances
General discoveryRitz-Carlton Orlando, Grande Lakes9/12
Luxury discoveryRitz-Carlton Orlando, Grande Lakes12/12
CouplesRitz-Carlton Orlando, Grande Lakes11/12
FamiliesJW Marriott Orlando, Grande Lakes10/12
BoutiqueThe Delaney Hotel12/12
SpaRitz-Carlton Orlando, Grande Lakes12/12
DiningFour Seasons Orlando11/12
ServiceRitz-Carlton Orlando, Grande Lakes11/12
Disney WorldWaldorf Astoria Orlando12/12
Water parksFour Seasons Orlando10/12

For the question, “What are the best boutique hotels in the Orlando area, Florida?”, The Delaney Hotel appeared in all twelve answers. For an upscale family vacation with children, JW Marriott Orlando, Grande Lakes led with ten appearances. For visiting Walt Disney World, Waldorf Astoria Orlando appeared in all twelve answers.

Four Seasons Orlando led the destination's combined total and was the most frequent answer for dining and the on-site water park question. That does not imply that each engine judged every recommendation an equally good fit. The counts record what the answers featured; they do not independently verify amenities, eligibility, location or suitability.

For hotel marketers, the practical distinction is between being visible for a broad destination question and being visible for a specific reason to travel. A boutique property can lead a relevant question without leading an aggregate dominated by several different travel needs.

AGR analyzed recommendation patterns for all 100 questions across the ten destinations. Orlando illustrates how those patterns can differ within a single market.

Repeating the same question often changed the hotel list

A single answer was rarely a complete description of an engine's behavior in this collection. Among 379 question-and-engine groups with three assessed answers, only 16 produced the same set of identified hotels in all three repetitions. The set changed at least once in 363 groups.

EngineQuestions with three assessed answersSame hotel set in all threeHotel set changed
ChatGPT91388
Google AI Mode91289
Gemini97394
Copilot100892

This comparison checks hotel membership, not wording or list order. A change may be as small as adding or removing one hotel while retaining the main recommendations. It does not mean that the entire answer changed or that a leading hotel disappeared.

The repetitions were collected on the same date. They show variation within this study, rather than a trend over weeks or months. They also explain why repeated observations are more informative than a screenshot of one answer.

Sources are part of the evidence, but not an explanation of causation

The saved answers contain hotel websites, review sites, travel publications and other links. AGR's analysis distinguished citation records from other attached links where the returned data made that distinction.

Follow-up research: AI Hotel Recommendations: Who Gets the Link? examines the sources and link destinations associated with the same 1,177 assessed answers collected on September 27, 2026. It distinguishes hotel and brand websites, publishers, listed sources, website buttons and Google travel interfaces, showing why a hotel recommendation does not necessarily include a link to the hotel’s own website.

A linked page is not automatically a recommendation of every hotel mentioned on that page. Likewise, a hotel displayed in ancillary data is not automatically a hotel featured in the answer. This study counts qualifying hotel entries in the answer itself, rather than treating all attached names or sources as recommendations.

Source visibility also does not establish why an engine selected a hotel. The study cannot determine the weight given to a hotel's website, reviews, an editorial article or any other input. The different engines expose sources in different ways, so this report does not rank them by source volume.

AGR's separate Luxury Hotel AI Recommendation Study examines which measured hotel attributes were associated with recommendation frequency among properties already recommended. That is a different question from what determines a hotel's initial inclusion.

For a hotel investigating its own results, the distinction between an AI visibility report and an AI visibility audit matters. A report records answer patterns. An audit also checks the hotel's public information for factual errors, inconsistent descriptions and gaps that may warrant correction. The appearance count alone does not diagnose the cause.

What hotel marketers can use from these findings

Measure the questions that match the hotel's guests. General discovery provides one view. Families, couples, spas, dining and local activities can reveal a different competitive set. The question-level results are often more useful than a single destination total.

Track engines separately as well as together. The nine destinations with differing engine leaders show how much a combined figure can conceal. A hotel's combined visibility should be accompanied by the underlying engine counts.

Use repeated observations before treating a change as a trend. One newly observed recommendation, or one absence, is insufficient evidence of a lasting gain or loss. Future editions can compare the same fixed questions while keeping each edition's date and coverage explicit.

Review the accuracy and clarity of the hotel's public information. Consistent property names and clear descriptions of location, facilities and guest suitability are sensible information practices. This study does not test whether making those changes increases AI recommendations, and it should not be presented as proof of an optimization formula.

That public-information work is the focus of AGR's Knowledge Formation Optimization (KFO) framework. KFO addresses how a hotel's identity is defined and supported across public sources, then measures whether AI answers describe and include it accurately over time. The research reported here measures recommendation behavior; it is not a test of a KFO intervention.

About the research

Americas Great Resorts assessed 1,177 substantive hotel recommendation answers collected on September 27, 2026, across ChatGPT, Google AI Mode, Gemini and Copilot. The fixed matrix comprises 100 questions across ten destinations, with three repetitions per engine. Tables show the assessed-answer denominator for each comparison.

This study measures which hotels were featured in AI answers; it does not measure how often the accompanying citations or links pointed to hotel websites, publishers or booking intermediaries.

Hotel names were normalized to consolidate clear spelling variants, confirmed rebrands and accommodation collections within the same hotel. Separately named hotels were retained separately; a joint featured entry credits each identifiable hotel once. Ambiguous brand-only references were not assigned to a specific property. Hotel names found only in source titles, ancillary metadata or incidental prose were excluded from the featured-appearance measure. AGR retained the original wording in its internal research records.

The destination results reflect the areas named in the questions and the properties the engines returned, including nearby properties. They are not a geographical certification or an endorsement. The engines' model versions, personalization and internal retrieval processes were not controlled uniformly. The findings apply to this collection and this prompt set, not every version or mode of these services.

Assess your hotel's AI visibility with AGR

For an individual hotel, the next step is to examine the questions that matter to its guests: whether the property appears, how it is described, which competitors recur, and whether the description matches the hotel's actual offering.

Americas Great Resorts provides a managed KFO service for independent luxury hotels and resorts. The work combines review and correction of the hotel's public information, development of supporting sources, and repeated measurement of AI descriptions and recommendation inclusion. Its purpose is to strengthen accurate representation; it does not guarantee placement in an AI answer.

Independent luxury properties can request an AGR AI Visibility Audit to examine their own recommendation presence, descriptions and competitive context, and identify priorities for further investigation or correction.

For related studies, practical guides and framework definitions, visit AGR's AI Visibility and KFO Resource Index.

Close