---
title: "AI Hotel Ranking — research & findings"
description: "Early findings from the AI Hotel Ranking census: how concentrated ChatGPT's hotel recommendations are, and where it gets its information for New York."
url: "https://sigtrip.com/ai-hotel-ranking/research/"
---

AI Hotel Ranking

# Research & findings

What the New York census tells us so far about how ChatGPT recommends hotels. These are our own measurements from our own runs; treat them as early signal, not settled fact.

## Recommendations are concentrated

Of 673 hotels tracked in New York, the engine names 368 across our questions — and citywide answers lean heavily on a small set of marquee names. On the overall board, the five most-named hotels alone account for about 29%of ChatGPT’s naming mentions. For most properties, the practical reality is simple: they never appear in the answer a traveler reads.

## The engine favors primary sources

When ChatGPT backs a New York recommendation with a citation, about 56.3%of those citations point to a hotel or brand’s own website, while only 2.9% point to travel-booking sites (OTAs like Booking or Expedia). For hotels, that makes an accurate, well-structured official site the single most influential asset for AI visibility. See the full breakdown on the [New York board](https://sigtrip.com/ai-hotel-ranking/new-york/) .

### Is your hotel on the board — or invisible?

See exactly where ChatGPT places you across real guest questions — then start improving it. Free scan, no credit card.

Start tracking my hotel →

or run a one-time free scan