{"@context":"https://schema.org","@type":"BlogPosting","headline":"AI Cross-Platform Consensus 2026: Five Engines, Five Lists","description":"A cross-engine comparison of AI hotel recommendations in one pinned week (2026-08-03, identical prompts on every engine): the per-city top-10 lists of ChatGPT, Gemini, Perplexity, Copilot and Google AI Mode intersected across the 56 destinations of the AI Hotel Landscape, hotels entity-resolved to canonical IDs and assigned to cities by geo bounding box. Of 1,490 distinct (city, hotel) recommendations, 57.6% exist on exactly one engine and only 4.6% (69 hotels) are backed by all five; the average recommendation has 1.83 engines behind it. Pairwise, Copilot × Google AI Mode is the most similar pair (34.8 avg Jaccard, 51.2 containment — the only pair sharing more than half its picks) and ChatGPT × Perplexity the least (16.4). Perplexity is the contrarian: 45.8% of its top-10 picks (252 of 550) appear on no other engine. Chain-branded hotels average 2.06 backing engines vs 1.65 for independents and reach all-five consensus at 6.6% vs 3.2% — a brand flag roughly doubles cross-engine agreement. City consensus runs from Auckland (49.6 avg pairwise Jaccard, 5 all-five hotels) to Bali (3.8, 44 distinct hotels across 50 top-10 slots, zero all-five); 17 of 56 cities — including Paris, London, Rome, Tokyo, Bangkok and Bali — have no all-five hotel. Exactly one hotel is ranked #1 on all five engines: The Taj Mahal Palace, Mumbai (1-1-1-1-1); Hotel Monteleone, New Orleans is the runner-up (#1 on four, #2 on ChatGPT). Grok excluded: data stale since week 2026-05-18, scraper suspended 2026-05-25.","datePublished":"2026-08-04","dateModified":"2026-08-04","url":"https://nicolassitter.com/research/ai-cross-platform-consensus-2026","category":"research","keywords":["AI cross-platform consensus","do AI engines agree on hotels","ChatGPT vs Gemini hotel recommendations","Perplexity hotel picks","AI hotel recommendations comparison","AI engine overlap","chain vs independent hotels AI","multi-engine AI visibility"],"articleSection":"Research","wordCount":2400,"readTime":"10 min","articleBody":"August 20265 engines × 56 cities, one pinned week\n\n# AI Cross-Platform Consensus 2026:five engines, five different hotel lists\n\nIf you ask ChatGPT, Gemini, Perplexity, Copilot and Google AI Mode which hotels to book in the same city, do they hand you the same names? I froze one week of the [AI Hotel Landscape](/projects/ai-hotel-landscape) data (2026-08-03, identical prompts on every engine) and intersected the five top-10 lists for each of the 56 destinations. The short answer: they barely overlap.\n\n**TL;DR.** Of the 1,490 distinct (city, hotel) top-10 recommendations that week, **57.6% exist on exactly one engine** and just 4.6% (69 hotels) are backed by all five. Copilot and Google AI Mode resemble each other most, ChatGPT and Perplexity least, and Perplexity alone sources 45.8% of its picks from hotels no other engine surfaces. Chain-branded properties collect roughly twice the all-five consensus of independents (6.6% vs 3.2%). And across every city and every engine, a grand total of **one** hotel holds the #1 slot everywhere: The Taj Mahal Palace, Mumbai.\n\n### Summarize with AI\n\n57.6%\n\nof recommendations are single-engine\n\n858 of 1,490\n\n4.6%\n\ncarried by all 5 engines\n\n69 consensus hotels\n\n45.8%\n\nof Perplexity picks are its alone\n\nhighest uniqueness of any engine\n\n1\n\nhotel ranked #1 on all 5\n\nThe Taj Mahal Palace, Mumbai\n\nThis closes a triptych. The [rankings-consistency study](/research/ai-hotel-rankings-consistency-study-2026) measured how much a single engine disagrees with _itself_ minutes apart (50.5% position-1 stability); the [volatility study](/research/ai-hotel-volatility-2026) measured how much it disagrees with itself week to week (29.2% top-10 retention). The remaining axis is engines disagreeing with _each other_ at the same moment on the same questions — and it turns out to be the widest gap of the three.\n\nAn average recommendation that week was backed by 1.83 of the five engines. In practice that means a traveler who switches assistants gets a mostly new shortlist, and a hotel that monitors only ChatGPT is blind to the majority of its AI surface area. The rest of this piece walks through where the agreement concentrates: which engine pairs, which hotel types, which cities.\n\n## 1\\. The setup\n\nThe landscape pipeline puts the same 616-prompt library (56 destinations × 11 templates) to each engine every week and entity-resolves every recommended hotel to a canonical ID. For this study I pinned the week of 2026-08-03 — the same week for all five engines, asserted by the analysis script — assigned each hotel to its destination by geo bounding box, and took each engine’s ten most-mentioned hotels per city. That yields up to 280 city-engine lists (26 come up shallower than ten hotels upstream), which flatten into 1,490 distinct (city, hotel) recommendations to intersect.\n\n-   **“Agreement” is set membership**, deliberately generous: two engines agree on a hotel if both place it anywhere in their top 10 for that city, regardless of position. The rank-level bar is much higher — section 7 shows what survives it.\n-   **Grok sat this one out.** Its most recent usable data is the week of 2026-05-18, eleven weeks before the pinned week (the scraper has been suspended since 2026-05-25). Comparing a May snapshot against August ones would manufacture disagreement, so the panel is the five engines with current data. The freshness table in the methodology documents this.\n-   **One week is a snapshot of a moving target.** Per the volatility study, an engine keeps only ~29.2% of its own top 10 from one Monday to the next, so the specific hotels named here will rotate. The disagreement _level_ is the durable finding; treat the hotel names as examples from 2026-08-03.\n\n## 2\\. The consensus curve\n\nBucket every recommendation by how many engines make it and the distribution collapses toward one: 858 of the 1,490 city-hotel picks (57.6%) live on a single engine’s list. Each additional engine roughly halves the bucket — 18.9% shared by two, 11.2% by three, 7.7% by four — down to the 69 recommendations (4.6%) that all five engines endorse.\n\nai-cross-platform-consensus-curve-2026\n\nThe five buckets partition all 1,490 distinct recommendations of week 2026-08-03 (shares sum to 100%).\n\nEngines agreeing\n\nCity-hotel recommendations\n\n% of all distinct\n\n1\n\n858\n\n57.6%\n\n2\n\n282\n\n18.9%\n\n3\n\n167\n\n11.2%\n\n4\n\n114\n\n7.7%\n\nAll 5\n\n69\n\n4.6%\n\nRead this as a market-structure fact: an “AI-recommended hotel” usually means **recommended by one AI**. Any claim of AI visibility that doesn’t name the engine is underspecified — 57.6% of these picks would evaporate if you checked a different assistant.\n\n## 3\\. Who agrees with whom\n\nPairwise, the disagreement has a clear geometry. The closest pair is **Copilot × Google AI Mode** at 34.8 average Jaccard — the only pair sharing more than half of each other’s picks by the containment measure (51.2%). Both lean on Bing/Google-flavored web retrieval, and it shows. At the far end sits **ChatGPT × Perplexity** at 16.4 — two chat assistants whose hotel taste has less in common than any other combination. Gemini pairs most naturally with AI Mode (31.2), which is intuitive for two Google surfaces, though even that overlap is under a third.\n\nAvg Jaccard %\n\nChatGPT\n\nGemini\n\nPerplexity\n\nCopilot\n\nGoogle AI Mode\n\nChatGPT\n\n—\n\n29\n\n16.4\n\n27.4\n\n29.9\n\nGemini\n\n29\n\n—\n\n19.4\n\n27.3\n\n31.2\n\nPerplexity\n\n16.4\n\n19.4\n\n—\n\n24\n\n22.2\n\nCopilot\n\n27.4\n\n27.3\n\n24\n\n—\n\n34.8\n\nGoogle AI Mode\n\n29.9\n\n31.2\n\n22.2\n\n34.8\n\n—\n\nJaccard similarity of two engines’ per-city top-10 hotel sets (intersection ÷ union), averaged over all 56 cities. 100 = identical lists, 0 = nothing in common. Darker = more agreement. These are pairwise similarities, so they don’t sum to anything.\n\n## 4\\. The contrarian\n\nAveraging each engine’s Jaccard against the other four turns the matrix into a personality ranking, and Perplexity finishes last by a wide margin: 252 of its 550 top-10 picks (45.8%) appear on _no other engine’s_ list for the same city. ChatGPT is a distant second at 33.4%; Google AI Mode is the most consensual engine at 22.7%. This rhymes with what the volatility data showed about Perplexity’s relationship with its own past picks — the engine that re-rolls its list weekly also strays furthest from the pack at any given moment.\n\nai-cross-platform-consensus-uniqueness-2026\n\nPer-engine totals across all 56 cities. 'Unique' = the (city, hotel) pick appears in no other engine's top 10 for that city.\n\nEngine\n\nTop-10 picks\n\nUnique to it\n\nUniqueness\n\nAvg Jaccard vs others\n\nPerplexity\n\n550\n\n252\n\n45.8%\n\n20.5\n\nChatGPT\n\n548\n\n183\n\n33.4%\n\n25.7\n\nGemini\n\n539\n\n167\n\n31%\n\n26.7\n\nCopilot\n\n544\n\n133\n\n24.4%\n\n28.4\n\nGoogle AI Mode\n\n543\n\n123\n\n22.7%\n\n29.5\n\nWhether Perplexity’s 45.8% is a bug or a feature depends on which side of the counter you stand. A traveler gets genuine variety by asking it second; a hotel that shows up only in Perplexity should know it is standing on the panel’s least corroborated — and, per the volatility study, least stable — surface.\n\n## 5\\. Chains vs independents\n\nSplit the 1,490 recommendations by whether the hotel carries a chain brand and the consensus concentrates visibly. A chain-branded pick is backed by 2.06 engines on average against 1.65 for an independent; half of chain picks (50.5%) have at least one corroborating engine versus 36.4% for independents; and at the all-five bar the gap is 6.6% to 3.2% — call it 2.1×. Independents supply most of the raw recommendations (850 of 1,490) but they overwhelmingly populate the single-engine tail.\n\nShare columns are cumulative (≥2 includes ≥3 includes all-5), so they intentionally overlap rather than sum to 100.\n\nSegment\n\nRecommendations\n\nAvg engines/rec\n\nShared by ≥2\n\nShared by ≥3\n\nAll 5\n\nChain-branded\n\n640\n\n2.06\n\n50.5%\n\n31.6%\n\n6.6%\n\nIndependent\n\n850\n\n1.65\n\n36.4%\n\n17.4%\n\n3.2%\n\nA brand flag is the closest thing to portable AI visibility this dataset shows. My working explanation is corpus redundancy — chains are documented consistently across the many different sources these engines retrieve from, while an independent’s visibility often hangs on whichever listicle or forum thread one particular engine happens to favor. Cause is untested here either way; the 2.1× consensus gap is what the week measured.\n\n## 6\\. The city gradient\n\nAveraging the ten pairwise Jaccards per destination spreads the 56 cities across a 13× range. Auckland tops the table at 49.6%: its five top-10s contain just 19 distinct hotels, 5 of which every engine lists. Cairo (46.9), Singapore (42.0), Queenstown (39.9) and Sydney (38.9) follow — compact markets where one obvious luxury shortlist exists and every engine finds it. At the bottom sits Bali at 3.8%: fifty top-10 slots holding 44 different hotels, with not a single property on all five lists. Phuket (10.3), Paris (11.8) and Prague (13.2) keep it company — huge, fragmented leisure markets where each engine assembles its own canon.\n\nThe number that surprised me most: **17 of the 56 cities have no all-five hotel at all**, and the list reads like a greatest-hits of world tourism — Paris, London, Rome, Tokyo, Bangkok and Bali are all on it. The deeper and more written-about a market, the less the engines can settle on a shared answer for it.\n\nAll 56 destinations, week 2026-08-03, sorted by consensus. 'Distinct hotels' counts the union of the five top-10 lists (50 slots when all lists are full).\n\nCity\n\nAvg pairwise Jaccard %\n\nDistinct hotels in 5 top-10s\n\nAll-5 hotels\n\nMost-agreed hotel\n\nEngines\n\nAuckland\n\n49.6\n\n19\n\n5\n\nPark Hyatt Auckland\n\n5/5\n\nCairo\n\n46.9\n\n17\n\n2\n\nFour Seasons Hotel Cairo at Nile Plaza\n\n5/5\n\nSingapore\n\n42\n\n21\n\n2\n\nMarina Bay Sands Singapore\n\n5/5\n\nQueenstown\n\n39.9\n\n20\n\n3\n\nEichardt's Private Hotel\n\n5/5\n\nSydney\n\n38.9\n\n24\n\n3\n\nFour Seasons Hotel Sydney\n\n5/5\n\nGold Coast\n\n38.4\n\n21\n\n3\n\nJW Marriott Gold Coast Resort & Spa\n\n5/5\n\nBuenos Aires\n\n37.4\n\n23\n\n3\n\nPalacio Duhau - Park Hyatt Buenos Aires\n\n5/5\n\nNairobi\n\n35.3\n\n23\n\n2\n\nSankara Nairobi, Autograph Collection\n\n5/5\n\nAmalfi Coast\n\n34.2\n\n25\n\n3\n\nSanta Caterina Hotel\n\n5/5\n\nBerlin\n\n34.1\n\n24\n\n3\n\nHotel Adlon Kempinski Berlin\n\n5/5\n\n## 7\\. The 69 hotels everyone agrees on\n\nSo who actually clears the bar? The 69 all-five hotels are spread across 39 cities, and 42 of them carry a chain brand. Then there is the question the whole study builds to: does any hotel hold rank #1 on every engine? Exactly one does. **The Taj Mahal Palace in Mumbai** is the most-mentioned hotel for its city on ChatGPT, Gemini, Perplexity, Copilot and Google AI Mode alike — a clean 1-1-1-1-1 in a dataset where five-way agreement on anything is rare. Volatility readers may recognize it: it was also one of the few hotels that never left a top-100 on two engines across that study’s eleven weeks.\n\nThe nearest misses are instructive too. **Hotel Monteleone** in New Orleans goes 2-1-1-1-1, denied unanimity only by ChatGPT’s #2. Hyatt Regency Cape Town (1-1-3-1-1) and Mandarin Oriental Wangfujing Beijing (1-3-1-1-1) each take four firsts, and Electra Palace Athens opens with three. Household names appear further down — Marina Bay Sands, Copacabana Palace, Royal Mansour Marrakech — but ranked differently everywhere, which is the study in miniature.\n\nEvery hotel in all five engines' top 10 for its city (week 2026-08-03), with its rank on each engine.\n\nCity\n\nHotel\n\nBrand\n\nChatGPT\n\nGemini\n\nPerplexity\n\nCopilot\n\nAI Mode\n\nAmalfi Coast\n\nAnantara Convento di Amalfi Grand Hotel\n\nMinor\n\n#1\n\n#3\n\n#1\n\n#4\n\n#2\n\nAmalfi Coast\n\nHotel Marina Riviera\n\nindependent\n\n#4\n\n#5\n\n#9\n\n#2\n\n#5\n\nAmalfi Coast\n\nSanta Caterina Hotel\n\nindependent\n\n#5\n\n#2\n\n#5\n\n#1\n\n#1\n\nAthens\n\nElectra Palace Athens\n\nindependent\n\n#4\n\n#2\n\n#1\n\n#1\n\n#1\n\nAuckland\n\nCordis, Auckland\n\nindependent\n\n#2\n\n#3\n\n#2\n\n#4\n\n#1\n\nAuckland\n\nInterContinental Auckland by IHG\n\nIHG\n\n#4\n\n#5\n\n#6\n\n#5\n\n#5\n\nAuckland\n\nPark Hyatt Auckland\n\nHyatt\n\n#1\n\n#1\n\n#7\n\n#3\n\n#2\n\nAuckland\n\nSofitel Auckland Viaduct Harbour\n\nAccor\n\n#7\n\n#2\n\n#1\n\n#1\n\n#3\n\nAuckland\n\nThe Hotel Britomart\n\nindependent\n\n#3\n\n#6\n\n#8\n\n#8\n\n#4\n\nBarcelona\n\nHotel Ohla Barcelona\n\nindependent\n\n#4\n\n#5\n\n#7\n\n#6\n\n#3\n\n## 8\\. What it means\n\n-   **For hotels: audit all five engines before believing any of them.** With 57.6% of recommendations single-engine, your ChatGPT visibility says almost nothing about your Gemini visibility. The pairs to treat as semi-redundant are Copilot & AI Mode (34.8 Jaccard) and Gemini & AI Mode (31.2); Perplexity needs its own check every time.\n-   **Corroboration is the quality signal.** A pick backed by three or more engines (23.5% of the total) sits in a different tier from a single-engine special — especially a Perplexity-only one, given its 45.8% uniqueness rate.\n-   **For travelers, a second assistant is nearly a second opinion by construction.** The average overlap between two engines’ top-10 lists for the same city is 26.2% (mean pairwise Jaccard across the ten pairs). If you want variety, switch engines; if you want reliability, look for the hotels they name in common.\n-   **Independents compete engine by engine.** The consensus layer is 42/69 chain-branded; an independent’s realistic goal in 2026 is being the darling of one or two engines and knowing which ones they are.\n-   **In fragmented markets, city-level AI rank is close to meaningless.** A Bali or Paris hotel can be #1 on one engine and invisible on four. In Auckland-type markets the shortlist is shared, and cracking it means displacing a hotel all five engines already agree on.\n\n## Methodology\n\n**Data:** the weekly AI Hotel Landscape pipeline — 616 prompts (56 destinations × 11 templates, EN, mixed personas/budgets) fired at ChatGPT, Gemini, Perplexity, Copilot and Google AI Mode; answers entity-resolved to canonical hotel IDs. This study reads the public dashboard’s geo view for the single pinned week of 2026-08-03 (6,613 rows; the analysis script asserts all five engines are on that exact week). Hotels are assigned to destinations by the dashboard’s geo bounding boxes: 0 rows dropped for missing canonical IDs, 41 rows (0.6%) fell outside every box and were excluded, none matched multiple boxes. Gemini figures use the `gemini_scraper` pipeline label — the dashboard’s “Gemini”.\n\n**Definitions:** an engine’s “top 10” for a city = its ten most-mentioned hotels there that week. Agreement between engines = shared membership of those per-city sets; Jaccard = intersection over union, containment = intersection over the smaller set, both averaged across the 56 cities. The consensus curve buckets each distinct (city, hotel) pair by how many engines’ sets contain it. Chain attribution comes from the landscape’s brand mapping.\n\n**Grok:** excluded because its data stopped at the week of 2026-05-18 (scraper suspended 2026-05-25 after persistent Class-C errors), while the five included engines all have rows for 2026-08-03:\n\nPlatform freshness at analysis time. The five current engines form the panel; Grok's 11-week-old snapshot is left out.\n\nPlatform\n\nLatest week\n\nRows that week\n\nVs pinned week\n\nStatus\n\nChatGPT\n\n2026-08-03\n\n1,775\n\ncurrent\n\nlive\n\nGemini\n\n2026-08-03\n\n1,182\n\ncurrent\n\nlive\n\nPerplexity\n\n2026-08-03\n\n1,251\n\ncurrent\n\nlive\n\nCopilot\n\n2026-08-03\n\n1,129\n\ncurrent\n\nlive\n\nGoogle AI Mode\n\n2026-08-03\n\n1,276\n\ncurrent\n\nlive\n\nGrok\n\n2026-05-18\n\n1,323\n\n11 weeks behind\n\nstale\n\n**Caveats:** the upstream table keeps each platform-week’s global top 500 plus a per-country floor of 50, and 13 of 206 platform-country groups show visible truncation — a bias that can only understate uniqueness and overlap gaps, never fabricate consensus. 26 of 280 city-engine lists surface fewer than ten hotels. Upstream unresolved-mention rates run 1.5–4.3% by platform (Perplexity highest at 4.3%). All of this describes English prompts in one pinned week; the volatility study quantifies how much such a week moves. Full per-table CSVs and headline stats: [summary.csv](/data/ai-cross-platform-consensus-2026/summary.csv) and the eight consensus\\_\\*.csv files in the same folder, generated by the landscape repo’s `analyze_consensus.py` (merged as ai-scrapers PR #41, with shape tests).\n\n## FAQ\n\n[← All research](/research)","author":{"@type":"Person","name":"Nicolas Sitter","url":"https://nicolassitter.com/about","sameAs":["https://www.linkedin.com/in/nicolassitternolleau/","https://github.com/Nicositter88","https://hotelrank.ai"]},"publisher":{"@type":"Person","name":"Nicolas Sitter","url":"https://nicolassitter.com"},"image":"https://nicolassitter.com/api/og/ai-cross-platform-consensus-2026","mainEntityOfPage":{"@type":"WebPage","@id":"https://nicolassitter.com/research/ai-cross-platform-consensus-2026"},"tags":["AI Search","Hotels","Consensus","Cross-Platform","AI Visibility"],"sameAs":["https://hotelrank.ai/research/ai-cross-platform-consensus-2026"],"alternateFormat":{"html":"https://nicolassitter.com/research/ai-cross-platform-consensus-2026","json":"https://nicolassitter.com/api/post/ai-cross-platform-consensus-2026","rss":"https://nicolassitter.com/rss.xml"},"datasets":[{"name":"summary","contentUrl":"https://nicolassitter.com/data/ai-cross-platform-consensus-2026/summary.csv","encodingFormat":"text/csv"}]}