AI Hotel Landscape
How leading AI assistants recommend hotels: 616 prompts across 56 destinations, a monthly edition checked before it is published.
+ 44 more
The 616 prompts we ask Copilot each edition
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 edition.
- 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 — Copilot
Hotels Copilot talked about most in this edition, with movement against the previous edition. Filtered to Copilot only — switch tabs to compare across models.
Where Copilot pulls hotel information from
Every URL Copilot retrieved classified into 12 buckets: OTAs, editorial, chain sites, direct hotel pages, social, meta-search, AI tools, community, government, directory consortia, and other.
Share of citations by source type, week by week. Search engine is its own bucket rather than part of metasearch, and covers Google (every regional domain), Bing and the rest. Google AI Mode routes its cited sources through google.com/searchviewer instead of linking the publisher, and those citations were landing under “meta-search” alongside trivago and kayak — which made the mix read as competitive pressure from metasearch when actual metasearch is under 1%. The dashed rule marks the week that started.
Weekly until 19 Oct 2026, monthly from November 2026
Which categories get a line is decided once over the whole window by total citations, not week by week — otherwise a category dropping out of the top N would look like it collapsed to zero. Everything below the cut is summed into “smaller categories” rather than dropped, so the lines always account for the full week. Citations per week: 05-11 3,348 · 05-18 3,669 · 05-25 3,592 · 06-01 3,924 · 06-08 3,659 · 06-15 2,916 · 06-22 3,123 · 06-29 2,746 · 07-06 2,668 · 07-13 3,386 · 07-20 5,637 · 07-27 5,702 · 08-03 4,039 · 08-10 4,513 · 08-17 4,310 · 08-24 4,275 · 08-31 9,154 · 09-07 4,210 · 09-14 4,604 · 09-21 4,756 · 09-28 4,782 · 10-05 4,828.
Top hotel parent groups by Copilot mentions
Each Copilot mention matched against Google Places, then rolled up to parent group (Marriott = Ritz-Carlton + Westin + Sheraton + EDITION + …, etc.). Change vs the previous edition. Detection via 175+ brand-domain rules.
| 1 | Wyndham | 14 → 0 | -100% |
| 2 | Taj | 16 → 0 | -100% |
| 3 | Shangri-La | 20 → 0 | -100% |
| 4 | Rosewood | 10 → 0 | -100% |
| 5 | Radisson | 23 → 0 | -100% |
| # | Brand | Mentions | Hotels | vs prev. |
|---|
What every metric means
Plain-English definitions for each number on this page. For deeper visuals + examples, see the annual landscape report linked at the bottom.
- Captures
- Number of AI responses we collected for this edition. Target = 616 per platform (one per prompt). Less when there are upstream errors.
- Web search
- Did the response trigger a live web fetch? Detected from the underlying response stream’s search-result events, not the unreliable top-level flag. We do NOT force web search — this is organic model behavior.
- Map widget
- Did the response render a hotel-card map (the Google-Maps-style widget ChatGPT shows for travel queries)? Each card is a "map entity" with its own provider.
- Sources / response
- Distinct URLs the model consulted while answering — the URLs that show up in its retrieval log, regardless of whether it cited them inline. Computed only over web-search responses (otherwise the answer is from training data, no sources to count).
- Citations / response
- Subset of sources that the model rendered as inline footnote pills in the answer. A source becomes a citation when the model explicitly references it.
- Fanouts / response
- Number of sub-queries the model spun up internally to answer the prompt (e.g. "best hotels Paris" might fan out to "hotels Paris Marais", "luxury hotels Paris", "rooftop hotels Paris", …).
- Map entities / response
- Hotel cards inside the map widget. Each carries a provider (Google Places, TripAdvisor, Yelp, Foursquare, SERP) and a place ID where applicable.
- OTA
- Online travel agency (Booking, Expedia, Hotels.com, Agoda, Trip.com, Priceline, etc.) — commission-based booking sites.
- Direct
- A hotel’s own website. Computed at request time by matching the cited domain against Google Places.
- Chain
- A hotel chain’s brand website (hilton.com, marriott.com, ihg.com, …). 175+ chain-domain rules.
- Editorial
- Travel media (Condé Nast Traveler, Time Out, Lonely Planet, Forbes Travel Guide, NYT, etc.) and city-specific travel blogs (santorinidave.com, theurbanlist.com, etc.).
- Directory
- Multi-property hotel consortia / brand collectives (Small Luxury Hotels of the World, Design Hotels, Preferred Hotels, Virtuoso, …). Aggregate hotels under one umbrella but aren’t a single OTA.
- Review
- TripAdvisor, Oyster, Yelp, Trustpilot, Holidaycheck — review-first platforms.
Does Copilot send people to Reddit and YouTube?
Share of answers citing at least one Reddit thread or YouTube video, week by week. Worldwide.
Share of answers citing at least one Reddit thread or YouTube video. Counted from the citation domains, so this measures what Copilot showed as a source — not whether it read them.
Weekly until 19 Oct 2026, monthly from November 2026
Latest week (2026-10-05): 0 of 1,232 answers cited Reddit (0.0%), 0 cited YouTube, across 4,828 citations.