AI Hotel Landscape
How leading AI assistants recommend hotels — 616 prompts across 56 destinations, refreshed every Monday.
THE OMNIA was the biggest mover with 17 mentions (+16 WoW). 2 hotels appeared in the top 200 for the first time.
ChatGPT names twice as many hotels as the next AI
For the same 616 prompts across 56 cities this week, ChatGPT mentioned **3,741 distinct hotels** by name — twice the next AI. Google AI Mode named about half (1,917), but cited **10 URLs per hotel** to back them. Two strategies coexist: breadth (ChatGPT) and source-density (AI Mode, Perplexity).
The 616 prompts we ask every week
56 destinations × 11 templates. Each prompt has structured dimensions: city, country, region, persona (couples / families / solo / business), budget (luxury / mid / budget), and location zoom (wide city vs neighborhood vs landmark). Control prompts are unmodified baselines.
Every destination below gets the same 11 question types — just with the city name swapped in. That's how 56 destinations × 11 templates = 616 prompts per platform per week.
- luxury hotels in <city>
- family friendly hotels in <city>
- hotels in <city> with rooftop pool
- hotels near <neighborhood>, <city>
- boutique hotels in <neighborhood>, <city>
- best hotels in <city> for solo travelers
- best hotels in <city>
- affordable hotels in <city> under $200
- best hotels in <city> for couples
- best hotels in <city> for business travelers
- best hotels in <neighborhood>, <city>
- hotels near <city> National Park
africa4 destinations · 44 prompts
americas14 destinations · 154 prompts
Each destination is asked the 11 prompt patterns above (11 prompts × 14 cities = 154 prompts in this region).
asia16 destinations · 176 prompts
europe17 destinations · 187 prompts
oceania5 destinations · 55 prompts
Each destination is asked the 11 prompt patterns above (11 prompts × 5 cities = 55 prompts in this region).
Named hotels of the week
Hotels worth calling out individually. Computed across all 6 AI assistants — when a hotel ranks high here it's mentioned by multiple models, not just one.
Most-mentioned hotel across all 616 prompts this week, blended across every AI assistant.
Wasn't on the radar last week. Newcomers gated to ≥15 mentions this week.
High-confidence cross-city detection lands with the May 18 run. We'll surface only mentions where the AI-named hotel exists in a different city than the prompt asked about — clean signal, no false positives from address formatting.
| # | Hotel | City | Country | ChatGPT | Gemini | Perplexity | Grok(discontinued) | Copilot | Google AI Mode | Total | Link mix |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | The Fullerton Bay Hotel Singapore | 80 Collyer Quay | SG | 13 | 8 | 2 | — | 4 | 7 | 34 | Direct 85%OTA 6%Other 9% |
| 2 | The Taj Mahal Palace, Mumbai | Mumbai | IN | 6 | 6 | 6 | — | 4 | 8 | 30 | Direct 97%OTA 3% |
| 3 | Four Seasons Hotel Sydney | The Rocks | AU | 6 | 5 | 7 | — | 7 | 4 | 29 | Direct 86%OTA 10%Other 3% |
| 4 | Hotel Monteleone | New Orleans | US | 6 | 7 | 3 | — | 6 | 7 | 29 | Direct 90%OTA 3%Other 7% |
| 5 | Cairo Marriott Hotel | Zamalek | EG | 12 | 8 | 2 | — | 1 | 4 | 27 | Direct 81%OTA 15%Other 4% |
| 6 | Peridot Grand Luxury Boutique Hotel | Ha Noi | VN | 5 | 7 | 4 | — | 4 | 7 | 27 | Direct 26%OTA 67%Other 7% |
| 7 | The Langham Gold Coast | Surfers Paradise | AU | 7 | 3 | 5 | — | 6 | 6 | 27 | Direct 89%OTA 4%Other 7% |
| 8 | Marina Bay Sands Singapore | 10 Bayfront Ave | SG | 6 | 5 | 3 | — | 7 | 5 | 26 | Direct 92%OTA 8% |
| 9 | Mandarin Oriental Wangfujing Beijing | Dongcheng | CN | 5 | 3 | 7 | — | 5 | 6 | 26 | Direct 88%Other 12% |
| 10 | Park Hyatt Auckland | Auckland | NZ | 6 | 4 | 6 | — | 5 | 5 | 26 | Direct 85%Other 15% |
| 11 | The Oberoi, Mumbai | Mumbai | IN | 7 | 5 | 3 | — | 3 | 6 | 24 | Direct 100% |
| 12 | Shangri-La Sydney | The Rocks | AU | 5 | 6 | 6 | — | 4 | 3 | 24 | Direct 79%OTA 17%Other 4% |
| 13 | QT Gold Coast | Surfers Paradise | AU | 6 | 7 | — | — | 5 | 6 | 24 | Direct 75%OTA 4%Other 21% |
| 14 | Hotel Grande Bretagne, a Luxury Collection Hotel | Athens | GR | 5 | 4 | 6 | — | 4 | 4 | 23 | Direct 83%OTA 4%Other 13% |
| 15 | The Oriental Jade Hotel | Ha Noi | VN | 4 | 7 | 5 | — | 3 | 4 | 23 | Direct 52%OTA 30%Other 17% |
| 16 | Mount Nelson, A Belmond Hotel, Cape Town | Gardens, Cape Town | ZA | 5 | 5 | 3 | — | 4 | 6 | 23 | Direct 87%OTA 4%Other 9% |
| 17 | The Venetian Las Vegas | Las Vegas | US | 4 | 6 | 3 | — | 5 | 5 | 23 | Direct 96%Other 4% |
| 18 | Park Hyatt Sydney | The Rocks | AU | 6 | 2 | 5 | — | 6 | 4 | 23 | Direct 96%OTA 4% |
| 19 | Santa Caterina Hotel | Amalfi | IT | 4 | 5 | 6 | — | 4 | 4 | 23 | Direct 87%OTA 9%Other 4% |
| 20 | QT Queenstown | Queenstown | NZ | 4 | 3 | 5 | — | 7 | 3 | 22 | Direct 86%OTA 9%Other 5% |
Link mix over time — top hotels
Where each AI sent users in recent weeks: direct site, OTA, chain page, or other.Direct vs OTA — who controls the link?
When an AI hands a user a hotel link, does it send them to the brand (own site or chain page) or to an OTA reseller? Bars fold chain into direct.How the 6 AI assistants compare
Same prompts, same week — different answers. Each model's top hotel, brand mix, and source preferences side by side.
| Assistant | Top hotel | City | Chain | Mentions | |
|---|---|---|---|---|---|
| ChatGPT | The Fullerton Bay Hotel Singapore | 80 Collyer Quay | — | 13 | deep dive → |
| Gemini | Cairo Marriott Hotel | Zamalek | Marriott | 8 | deep dive → |
| Perplexity | Anantara Convento di Amalfi Grand Hotel | Amalfi | Minor | 8 | deep dive → |
| Grok(discontinued) | — | — | — | 0 | deep dive → |
| Copilot | Four Seasons Hotel Sydney | The Rocks | Four Seasons | 7 | deep dive → |
| Google AI Mode | The Taj Mahal Palace, Mumbai | Mumbai | Taj | 8 | deep dive → |
| Chain | ChatGPT | Gemini | Perplexity | Grok(discontinued) | Copilot | Google AI Mode | Avg |
|---|---|---|---|---|---|---|---|
| Marriott | 28.5% | 28.3% | 28.9% | — | 26.8% | 26.8% | 27.9% |
| Accor | 15.4% | 16.1% | 6.5% | — | 14.7% | 12.8% | 13.2% |
| Hilton | 11.7% | 9.3% | 17.2% | — | 12.5% | 12.1% | 12.6% |
| Hyatt | 10.6% | 10.4% | 10.5% | — | 9.1% | 9.7% | 10.1% |
| IHG | 9.2% | 9.2% | 12.7% | — | 9.9% | 8.7% | 9.9% |
| Four Seasons | 10.0% | 9.5% | 5.5% | — | 10.2% | 10.2% | 9.1% |
| Mandarin Oriental | 3.8% | 3.5% | 2.8% | — | 3.7% | 4.7% | 3.7% |
| Radisson | 2.0% | 3.1% | 6.2% | — | 3.0% | 3.4% | 3.4% |
| Minor | 2.4% | 2.0% | 3.8% | — | 2.6% | 3.2% | 2.8% |
| Shangri-La | 2.1% | 3.1% | 2.3% | — | 3.1% | 2.6% | 2.6% |
| Taj | 2.9% | 3.0% | 1.6% | — | 2.1% | 2.6% | 2.4% |
| Langham | 1.5% | 2.3% | 2.0% | — | 2.2% | 3.2% | 2.2% |
| Category | Top domains | Share | Citations |
|---|---|---|---|
| Meta-search | google.com(13,449) · kayak.com(297) · hotelscombined.com(119) | 15.4% | 14,200 |
| OTA | booking.com(1,498) · expedia.com(1,381) · hotels.com(621) | 15.3% | 6,049 |
| Other | fity.club(29) · hauteliving.com(28) · hanoitourism.org(23) | 15.1% | 5,194 |
| Independent | pestana.com(26) · scottdunn.com(18) · qthotels.com(17) | 14.4% | 5,181 |
| Chain | marriott.com(778) · hilton.com(413) · fourseasons.com(223) | 11.9% | 4,001 |
| Editorial | thehotelguru.com(559) · cntraveler.com(308) · timeout.com(266) | 10.3% | 3,513 |
| Review | tripadvisor.com(2,226) · tripadvisor.co.uk(302) · tripadvisor.ca(103) | 8.9% | 3,487 |
| Community | hotelierschoice.com(312) · myboutiquehotel.com(182) · santorinidave.com(146) | 4.6% | 1,420 |
| Directory | designhotels.com(64) · hotelsforkings.com(52) · boutiquehotel.me(23) | 1.4% | 450 |
| Social | youtube.com(404) · reddit.com(194) · facebook.com(154) | 1.2% | 949 |
| DMO | neworleans.com(41) · miamiandbeaches.com(29) · visitdublin.com(12) | 1.1% | 395 |
| AI Tool | staygenie.app(99) · wanderlog.com(2) · italyplanner.ai(1) | 0.3% | 104 |
| Encyclopedia | en.wikipedia.org(5) · en.wikivoyage.org(1) | 0.0% | 6 |
| Assistant | Mix (all categories) | Citations |
|---|---|---|
| ChatGPT | Review 25%OTA 20%Editorial 17%Other 17%Community 8%Meta-search 3%DMO 3%Chain 3%Directory 2%Independent 1%AI Tool 1%Social 0% Inside Other (17%): hanoitourism.org(23) · lucalvry.com(23) · blog.hotelslash.com(19) | 5,801 |
| Gemini | OTA 25%Other 24%Editorial 22%Community 11%Independent 6%Chain 4%Directory 3%Meta-search 2%DMO 1%Social 1%AI Tool 0%Review 0% Inside Other (24%): hauteliving.com(28) · hashcorner.com(21) · jetmeaway.co.uk(16) | 4,086 |
| Perplexity | Independent 27%OTA 22%Chain 17%Review 14%Other 9%Editorial 4%Meta-search 3%Community 2%Directory 0%DMO 0%AI Tool 0%Social 0% Inside Other (9%): discoverasr.com(20) · hotel-la.in(8) · cityparkhotel.in(7) | 10,858 |
| Grok(discontinued) | no citations captured | 0 |
| Copilot | Independent 36%Chain 34%Other 18%Review 4%Editorial 3%OTA 2%Community 1%Directory 1%Social 0%DMO 0%Meta-search 0% Inside Other (18%): fity.club(29) · travoh.com(18) · theluxevoyager.com(16) | 4,310 |
| Google AI Mode | Meta-search 68%Other 8%OTA 7%Editorial 5%Social 4%Independent 2%Review 2%Chain 2%Community 1%Directory 1%DMO 1%AI Tool 0%Encyclopedia 0% Inside Other (8%): foratravel.com(15) · secretplaces.com(11) · akbartravels.com(8) | 19,894 |
How AI hotel recommendations evolve
19 weeks of history. Lines aggregate across all 6 AI assistants.
Need a specific week's snapshot? Pick one from the weekly archive.
What's tracked + what's coming
The dashboard launches with ChatGPT. Five more models follow on the same cadence, plus deeper analysis on top.
- ✓ChatGPT — live now, 2,300+ captures every Monday across 56 destinations and 37 countries.
- ✓616 hotel-search prompts per platform per week — 11 question patterns × 56 cities (luxury, family, romantic, design, neighborhood, landmark, …).
- ✓Source mix — every URL ChatGPT cites is bucketed into OTA, editorial, chain site, direct hotel page, review, social, meta-search, AI tool, community, government, or directory.
- ✓Brand share — top hotel chains by mentions, week-over-week deltas.
- ✓Map widget — chain vs independent split, and where the map data comes from (Google Places, Yelp, TripAdvisor, …).
- ✓Sponsored placements — paid ad slots tracked separately from organic recommendations.
- Five more AI assistants
Gemini, Perplexity, Grok, Copilot, and Google AI Mode are already being scraped on the same weekly cadence. Their dashboards unlock as we validate the data — same depth as the ChatGPT view, plus side-by-side comparisons.
- AI Consistency Score
How stable is a hotel's ranking across the week, across queries, and across models? A single number per property that brand teams can track over time.
- Hotel of the Week + Hallucination of the Week
Recurring stories from the data: the property AI is currently obsessed with, and the funniest mistake it made (e.g. a hotel returned for the wrong city).
- Per-destination drilldown
Click any city to see exactly which hotels each AI recommends, which sources it cites, and how those choices have shifted week over week.
- Multi-week trend lines
Charts on every metric so you can see whether OTAs are gaining ground, which chains are climbing, and where ad density is heading.
How this dashboard is built
Full transparency on data collection.
What "mention" means
Hotel mention = a real-world hotel property that the AI named in its answer, after we matched the name against Google Places. A hotel mentioned multiple times in the same answer counts once.
Parent group mention = a hotel mention rolled up to its parent group via the chain taxonomy. A stay at the Ritz-Carlton counts toward Marriott; a stay at Sofitel counts toward Accor.
Citation / source = a URL the AI fetched from the web during the answer. Tracked separately from mentions. The Source mix charts answer "where does the AI get its information," not "which hotels does it recommend."
Link = the URL the AI gave specifically for a recommended hotel (its booking link). Classified as direct (hotel-owned), OTA (Booking, Expedia, …), chain page (Marriott.com, Hilton.com, …), or other.
Prompts. 616 hotel-search prompts, 56 destinations × 11 templates (e.g. "best hotels in Paris", "family friendly hotels in Cancún", "hotels near Big Ben").
Web search is NOT forced. We measure organic behavior. The web-search rate displayed here is derived from the underlying response stream's search markers, not the unreliable top-level flag.
Hotel matching. Each named hotel mention is matched against Google Places to resolve it to a real property (with address, website, and chain when applicable). The "Direct" source bucket is computed at request time against the same data.
Country. US proxy. As of May 2026, OpenAI's paid sponsor placements are rolling out in the US, Australia, New Zealand, and Canada — the test pool may expand. Geo coverage on this dashboard expanding to match.
Data is directional. Some prompts retry due to upstream variance; absolute counts can shift ±5% week to week. Trends are honest; single-week spot reads should be taken with that grain of salt.
Headline mover thresholds. Biggest Riser and Hotel That Disappeared require a prior-week base of ≥10 mentions; Newcomer requires ≥15 current-week mentions. Below these bases, % swings are noise (a 3 → 11 jump looks like "+267%" but carries no signal). Hotel of the Week is exempt — it's the absolute top, not a delta. Absolute counts are shown first, percentages second.
Refresh. Cron runs every Monday at 02:00 UTC. Page revalidates hourly.
Want a deeper read? See the in-depth annual report