20.3% of Hotels Receive Zero AI Recommendations; Only 17.2% Are 'Consistently Recommended'

Invisible Even to Luxury Giants: Hotel Search Structures in the AI Era and the Crossroads of 'Reducing OTA Dependency'

Terrace Roots, a consulting firm specializing in Generative Engine Optimization (GEO) for the hospitality industry, conducted a proprietary survey in May 2026 on the 'recommendation visibility' of 64 representative city, luxury, design, and upper-upscale business hotels across 8 major Japanese cities: Tokyo, Yokohama, Nagoya, Osaka, Kyoto, Sapporo, Sendai, and Fukuoka.

The study tested 6 queries per city, totaling 1,152 judgments across three generative AI services (ChatGPT, Perplexity, and Google AI Mode). Results show that 20.3% (13 properties) were never recommended by any of the three AI services. Only 17.2% were consistently recommended (recommendation rate of 40% or higher), meaning 82.8% of major hotels are effectively outside the AI recommendation loop.

### Background: Travel Planning Shifts from 'Search' to 'AI Consultation'

AI adoption in the travel industry is accelerating rapidly between 2025 and 2026. According to the US research firm Phocuswright, 56% of travelers used generative AI for planning, booking, or on-site usage in the past 12 months. In the domestic market, a survey by Shukuken (February 2026) revealed that 38.6% of generative AI users would have chosen standard, famous hotels if they hadn't used AI. This indicates that AI is pushing choices beyond 'standard' options, putting hotels that are not recommended at risk of losing their competitive standing.

### Four Structural Findings

1. **20.3% Receive Zero Recommendations:** Only 17.2% are consistently recommended, leaving 53 out of 64 properties effectively invisible. 2. **Gap Between Brand Strength and AI Visibility:** Some historic luxury hotels had a recommendation rate of only 11.1%, proving that traditional name recognition does not equate to AI visibility. 3. **Differences by Platform:** Google AI Mode had the highest 'zero recommendation' rate at 46.9%, illustrating a paradox where high SEO ranking does not guarantee AI recommendation. 4. **Severity in Regional Cities:** In some cities, the 'zero recommendation' rate reached 62.5%, highlighting significant disparities in 'discoverability' between locations.

### Management Impact: The Path to Reducing OTA Dependency

This survey result is not merely a marketing issue. For hotels struggling with OTA commissions (typically 15-30%), securing AI recommendations is directly linked to improving profit structures and reducing dependency on third-party booking platforms.