System for Measuring Semantic Trust Patterns in AI and Search Systems

🧠 Provisional Patent Overview

Title:
System for Measuring Semantic Trust Patterns in AI and Search Systems
📅 Filed: July 5, 2025 | 📄 Pages: 20
📎 Download PDF

---

🔍 What This Patent Covers

This invention introduces the EEAT Rank™ and AI TrustRank™ framework — a system that measures inferred trust in AI and search systems by analyzing how often a named entity appears in semantic proximity to high-authority sources.

Instead of backlinks or Schema, it uses co-occurrence detection across content formats like articles, transcripts, and FAQs to build a quantifiable trust score and graph-based trust map.

📊 Core Metrics & Concepts

Metric / Model	Purpose
EEAT Rank™	A 0–100 trust score based on co-occurrence patterns with authoritative sources (e.g., CMS.gov, Harvard.edu).
TrustRank™	An aggregated, domain-wide score computed from multiple EEAT Rank instances.
Trust Graph™	A visual and API-ready graph that maps how entities co-occur with trusted sources.
Trust Signal Weighting Engine	Weighs signal strength using proximity, authority, format diversity, and recurrence.

🧩 How It Works

FIG. 1–5 explain the full process:

1. Entity + Trusted Source Extraction
   Pulls named entities and matches them with high-authority domains.
2. Semantic Proximity Evaluation
   Scores strength of co-occurrence (sentence, paragraph, section).
3. Trust Signal Weighting
   Adds multipliers based on:
   • Distance
   • Authority of the source
   • Format diversity (blog + FAQ + podcast = stronger)
   • Temporal consistency (sustained, not bursty)
4. EEAT Rank Score Generation
   Normalizes signals into a trust score (0–100 scale or confidence bands).
5. Trust Graph Construction
   Outputs a machine-readable trust map with entity relationships.

🔬 Use Cases

• AI Search Engines: Prefer EEAT-ranked results when summarizing content.
• Publishers: Benchmark their trust footprint across industries.
• Compliance Tools: Detect risky entities lacking strong co-occurrence trust signals.
• TrustStacker Systems: Surface high-trust glossary pages, FAQs, and citations in AI Overviews.

🔐 Core Claims (Condensed)

• Claim 1: Full system claim: detects co-occurrence, weighs it, and outputs EEAT Rank + Trust Graph.
• Claim 4: Includes time and format weighting for durability and trust richness.
• Claim 7: EEAT Rank can be used in AI ranking or document retrieval.

📄 Full claims detailed on pages 17–18.

🔗 Related Glossary Terms

• EEAT Rank™
• Trust Graph™
• Trust Signal
• Semantic Proximity
• Co-Occurrence

🧠 Strategic Insight

This is the measurement layer that complements the other two patents:

• Patent #1 = TrustCast (the propagation engine)
• Patent #2 = AITO Feedback Loop (the conditioning method)
• Patent #3 = EEAT Rank + Trust Graph (the scoring + validation infrastructure)

Together, they form the full loop:
Propagate → Condition → Measure → Repeat

📎 Download Patent PDF

System for Measuring Semantic Trust Patterns (PDF, 20 pages)

Related

How Structure, not Schema, is Changing the AI Visibility Landscape