Domain Authority, Schema Markup e Geo Data: I tre livelli che l'IA usa per decidere se citare le tue pagine

Most guides to AI search ranking cover two layers: domain authority and schema markup. Those guides are not wrong, but they are incomplete in a way that specifically hurts listing pages, property portals, vacation rental platforms, and any site where inventory is location-based.

The third layer is geo data. It is the least documented, the most commonly missing, and the one that determines whether your pages can answer location-specific queries at all. Understanding what AEO actually means is the starting point, but this guide goes deeper into the structural factors that drive whether individual pages get cited.

Layer 1: Domain and Entity Authority

Domain authority is the entry requirement, not the ranking signal. Think of it as a threshold. Pages from domains below roughly DA 20 to 30 rarely appear in AI citation pools for competitive queries, regardless of content quality. Above that threshold, raw DA has a weakening correlation with citation frequency.

What replaced it as the primary signal above the DA floor is entity authority: how clearly and consistently AI models understand what your site is, what it covers, and who it serves.

Consistent entity identity across the web. Your organization name, address, URL, and category must appear identically across your own site schema, Google Business Profile, industry directories, and citation sources. NAP inconsistency directly fragments your entity identity across multiple weak representations instead of one strong one.

Topical coherence. AI models assess whether your site has a clear, consistent topic cluster. A site with 30 articles in one narrow niche is more entity-authoritative in that niche than a site with the same DA spread across 20 unrelated topics.

sameAs references. The sameAs property in your JSON-LD links your entity to its representations on Wikidata, Crunchbase, LinkedIn, and other authoritative graphs. AI models use these to confirm the entity they are reasoning about is the same one described across multiple sources. The complete LocalBusiness JSON-LD implementation guide covers how to structure this correctly.

If your domain clears the DA floor, entity authority improvements will do more for AI citation than additional link-building.

Layer 2: Schema Markup

Schema markup is the communication layer between your pages and AI retrieval systems. Pages with structured data are cited at significantly higher rates than unschemed pages. Google AI Overviews favor pages with structured data, and the selection lift is material for competitive queries.

Most implementations stop at the fields that satisfy Google's Rich Results Test, which is not the same as satisfying AI citation systems.

What most implementations get right: @type, name, description, url, openingHours, telephone, address, FAQ schema.

What most implementations miss for listing pages: The schema types designed for listing inventory require different properties from the types most guides discuss.

For real estate, vacation rental, and hospitality listing pages, the relevant types are RealEstateListing, LodgingBusiness, Hotel, VacationRental, Apartment, and SingleFamilyResidence, each nested with Offer for pricing and availability. These types only perform their function for AI retrieval when combined with the right location properties.

The FAQ Schema Mistake

FAQ schema is valuable for editorial content. It tells AI engines exactly which question a piece of content answers. Listing pages are not editorial content. A property listing is not answering a general question about vacation rentals. It is representing a specific entity at a specific location. FAQ schema does not help an AI engine match that listing to "2-bed apartment near the metro." The right schema for listing pages is entity-relational, not Q&A-shaped.

Layer 3: Geo Data (The Under-Documented Layer)

AI models that answer location-specific queries ("vacation rentals near Yellowstone," "apartments within 10 minutes of downtown") are doing implicit geospatial matching. They are resolving geographic relationships between the queried location and the entities in their retrieval pool. For that matching to work, your listing pages need to encode those relationships explicitly in structured data.

Precise GeoCoordinates on Every Listing Page

The GeoCoordinates geo property with latitude and longitude to at least four decimal places is the foundational signal. Without it, AI engines geocode your address string, which fails on any inconsistency and produces far lower precision. Most implementations that include geo at all apply it only to a site-level LocalBusiness schema, not to individual listing pages. Each listing page must be its own resolvable geographic entity.

"geo": {
  "@type": "GeoCoordinates",
  "latitude": 48.8566,
  "longitude": 2.3522
}

containedInPlace: Linking the Property to a Geographic Hierarchy

The containedInPlace property links your listing to the neighborhood, district, city, and region entities that contain it. This is how AI engines answer queries like "apartments in the Marais" rather than just "apartments at [street address]." Without it, a property exists as an address but not as a member of any geographic entity.

"containedInPlace": {
  "@type": "Place",
  "name": "Le Marais",
  "containedInPlace": {
    "@type": "City",
    "name": "Paris"
  }
}

Nearby Place Entities: Transit, Schools, Landmarks

When a user asks for "rentals near the metro," the AI is looking for explicit machine-readable relationships between the property and transit infrastructure. A sentence in your description saying "5 minute walk to Metro Line 4" does nothing for AI retrieval. The same information structured as a Place entity linked via amenityFeature is retrievable.

Why Listing Databases Do Not Carry This Data Natively

Most property management systems and listing databases store what operators enter: address, price, bedrooms, bathrooms, photos. They were built for humans browsing a portal, not for machine-readable geographic context. A mapping API fills this gap. Geocoding APIs convert addresses to precise coordinates. Points-of-interest APIs return transit stops, schools, parks, and landmarks within a given radius. The output maps directly to schema.org types and can be embedded into listing page JSON-LD at scale.

What Closing All Three Gaps Looks Like

A listing page that performs well in AI retrieval:

1. Lives on a domain with consistent entity identity, sameAs references, and a clear topical cluster
2. Uses the most specific applicable schema type nested with Offer for pricing
3. Includes GeoCoordinates on the listing page itself, containedInPlace linking it to neighborhood and city entities, and structured nearby Place data for transit, schools, and landmarks

Most listing pages cover parts of Layer 1 and basic parts of Layer 2. Almost none cover Layer 3. The pages that cover all three are the ones that appear in AI answers for location-specific queries.

Only 1.2% of local businesses currently appear in AI search recommendations. They are not, on average, the ones with the highest domain authority. They are the ones that have closed all three gaps.