SiteNexis models how AI systems retrieve, interpret, trust, and recommend web content — across twelve dimensions, sixteen autonomous agents, and every major AI surface.
Architecture
Each layer depends on the one below. Layer 4 systems cannot produce meaningful output without the entity graph, retrieval scores, and semantic signals from Layers 2 and 3.
Models how AI systems form, maintain, and lose trust in your content over time.
Scores citation probability, AI extractability, retrieval readiness, and recommendation confidence.
Extracts and scores every named entity, relationship, and semantic structure across your domain.
Full-site Puppeteer crawl: HTML parsing, chunk extraction, link graph, SEO signals, performance.
How AI Systems Read
Every step from raw HTML to AI recommendation is a potential failure point. SiteNexis instruments all eight stages.
Scoring System
Every score is 0–100. Every deduction maps to a named, actionable issue. No black boxes.
Composite of all AI-layer signals. The primary measure of how an AI system perceives your site.
Extraction fidelity — how much meaning survives the AI retrieval pipeline from raw HTML to usable chunk.
Scores entity detection, consistency across pages, coverage depth, and disambiguation strength.
Query-answer alignment, chunk extractability, and conversational query structure across 6 query types.
Likelihood an AI selects this content as a citation — factual density, claim specificity, authority depth.
Authorship, organisational, content, and structural trust signals — with cross-page contradiction detection.
Simulates chunk extraction, ranking pressure, summarisation loss, and citation eligibility filtering.
Entity credibility, schema alignment, external validation depth, and trust degradation resistance.
Rate of authority growth or decay across consecutive audit snapshots — velocity, not snapshot.
Inclusion probability across AI Overviews, chat, voice assistants, and autonomous agent discovery.
Detects synthetic entity patterns, manufactured authority networks, and schema manipulation signals.
Top-level composite. How deeply does an AI ecosystem trust, retrieve, and recommend this website?
Multi-Provider Intelligence
Provider scores are probabilistic estimates based on measurable content signals — not live API queries. Weights are configurable per provider.
| Provider | Primary Signal | Trust Mechanism | Citation Behaviour |
|---|---|---|---|
| Google AI Overviews | E-E-A-T + structured data | Domain authority + schema | Direct answer extraction |
| ChatGPT / GPT-4o | Semantic query embedding | Recency + factual density | Chunk-level quotation |
| Perplexity | Real-time crawl + ranking | Source diversity + directness | Inline citation with URL |
| Gemini | Knowledge Graph integration | Entity consistency + schema | Knowledge panel association |
| Claude | Semantic clarity + structure | Authoritativeness + entity clarity | Summarisation + attribution |
| Voice Assistants | Answer directness + schema | Structured data + authority | Single-answer extraction |
Run a free audit in under 5 minutes. No credit card. No install.