About the Observatory

Property discovery is moving from search pages with filters to AI conversations with natural language. This shift changes everything: how travelers find properties, which properties are seen, and how selection happens before a human ever views a listing.

Yet this transition is happening without transparency. Property operators cannot see whether their properties are being selected by AI systems, what signals matter, or how to improve their representation in AI-mediated discovery.

The Observatory exists to increase transparency. We observe how AI systems currently discover and select hospitality properties, identify patterns, and provide intelligence that helps operators understand the new landscape.

The Observatory is published by HomeSelf, but it is independent in its research methods and findings. HomeSelf provides the VPR (Verified Property Record) protocol and registry, which makes properties machine-readable for AI systems.

The Observatory observes how AI systems respond to hospitality queries and maps findings to VPR schema fields. This helps answer the question: "Which structured data elements correlate with property surfacing in AI responses?"

Importantly, the Observatory is not a marketing tool for HomeSelf. We observe patterns across all properties, regardless of whether they have VPRs. Our goal is transparency, not promotion.

VPR (Verified Property Record) is the HomeSelf protocol for publishing machine-readable property data. A VPR provides structured information about location, amenities, booking policies, and more in a format that AI systems can parse and evaluate.

The Observatory analyzes which VPR fields appear to correlate with property surfacing in AI responses. For example, if properties with detailed business amenity data surface more consistently for business travel queries, this suggests that VPR's amenity structure aligns with AI selection patterns.

The VPR Alignment Score in our benchmarks quantifies this correlation. A high score means observed conversational patterns map well to VPR schema; a lower score suggests opportunities for schema evolution or different representation approaches.