AGENTIC|

Research

AI Search & Digital Experiments

Data-driven studies on how AI search, the web, and digital platforms work. Currently spanning hotel-industry studies and cross-industry AI-search methodology. All research is open and independently published.

All Hotels AI Search Studies

LIVE · FeaturedHotels

AI Hotel Landscape — live weekly data

Refreshed every Monday from 616 prompts across 56 destinations. Brand leaderboards, source mix (OTA vs direct), per-AI breakdown, sponsored ads — watch how recommendations shift week over week across ChatGPT, Gemini, Perplexity, Grok, Copilot, and Google AI Mode.

616Prompts / week

MonRefresh cadence

Hotels

How AI models discover, rank, and recommend hotels.

The ChatGPT Direct-Traffic Explosion for Hotels

May 7, 2026 — daily AI sessions to hotels jumped +62% overnight

+62% Daily lift17K+ Hotels43 Get 10%+ from AI

The Schema.org Debate

Doesn't feed the LLM. Feeds what the LLM calls. Especially for hotels.

How Mistral Searches Hotels

One Brave call. Snippet paraphrase. No place data.

How Claude Searches Hotels

Default mode → Google Maps. Connectors on → curated OTA picker.

ChatGPT Hotel Ads Are Live

CPC pivot, ads manager, $50K entry — Booking.com still leads at 43.5%

27% Ad Rate43.5% Booking.com9 Advertisers

ChatGPT 5.3 Halved Its Hotel Sources

March 5, 2026 cutover — URLs/answer 24 → 12

140 Prompts24 → 12 URLs/answer132 → 61 Domain pool

Hotel YouTube Channels — Activity Study 2026

Only 10.0% of hotels have YouTube; 43.7% are ghost accounts

10.0% Hotels w/ YT43.7% Ghost11.3% Active

ChatGPT Hotel Index vs Live Web

83% of cited domains change between modes

400 Queries83% Domains Differ13 Zero Overlap

How Dirty Is Google Maps Hotel Data?

17% of 179K listings fail quality checks

179K Listings17% Fail QA11 Countries

ChatGPT Hotel Data Sources

Google fell from 100% to 70.3% in 90 days

100K Entities70.3% Google Share8.8x TA vs Google

What Hotel Footers Reveal

Instagram in 40.8%, stale copyrights in 24%

98,423 Hotels40.8% Instagram9.9% Link OTAs

Hotel llms.txt Adoption Study

Only 6.3% have a llms.txt file

105,002 Hotels6.3% Have llms.txt7 Countries

Hotel robots.txt & AI Blocking Study

Only 3.3% block any AI crawler

105,002 Hotels3.3% Block AI7 Countries

What Hotels Are Actually Called

A naming study of 121,425 hotels

121,425 Hotels8 Analyses7 Countries

Hotel Schema.org Adoption Study

36.3% of hotels have no schema at all

121,425 Hotels7 Countries36.3% No Schema

How Consistent Are AI Hotel Rankings?

Only 50.5% position stability across reruns

4,000 Queries50.5% Stability8 Cities

Anatomy of a ChatGPT Hotel Search

12 systems, 7 providers, 424 A/B tests

12 Systems7 Providers424 A/B Tests

Yelp in ChatGPT: Hotel Data Study

33% Yelp rate in US hotel queries

14 Destinations5 With Yelp33% Yelp Rate

Google AI Mode: Where Do Hotel Clicks Go?

79% of clicks go to Google Business Profiles

4,000 Queries84K+ References79% GBP Clicks

The AI Hotel Landscape 2026

How 6 AI Models Rank 12,500+ Hotels

1.2M+ Citations6 AI Models12,500+ Hotels

Do French Hotels Blog?

49.3% have blogs, only 1 in 4 are active

15,155 Hotels49.3% Have Blogs25% Active

In Progress

Coming Soon

The Palace Bias

62% palace share in 5-star AI recommendations

Coming Soon

Hotel Page Speed Study

Core Web Vitals audit

Coming Soon

Location & Language Bias

2 languages, US vs FR compared

How to Measure AI Hotel Traffic and Bookings

GA4 referrers, branded queries, booking-form prompts, ChatGPT Apps

AI Fan-Out Query Anatomy

22,518 AI hotel sub-queries decoded

Trust in AI hotel answers

Cross-checked across sources — but mostly anchored on Google Maps

The "fake 1-star" hotel problem

When Google Maps gets the category wrong, AI inherits the error