About SitePulsar
How our AI audit engine works and what drives each score.
What SitePulsar does
SitePulsar audits how well a website is positioned in the age of generative AI. It runs an Agent Experience Optimization (AEO) audit with three pillars: FIND — brand visibility inside AI engines (ChatGPT, Claude, Gemini, Perplexity, Grok) plus on-site discoverability; READ — schema depth, structured data, trust signals, and content comprehension; and USE — how ready the website is for autonomous AI agents that browse, extract data, and take actions on behalf of users.
Every audit is generated by AI and every report carries an AI-generated disclaimer in compliance with the EU AI Act (Article 50). Scores are informational — not a guarantee of ranking or agent compatibility.
The Penta-Core scoring engine
Our Penta-Core engine maps eleven sub-signals into three AEO pillars (FIND, READ, USE). Scores are calculated deterministically from the underlying sub-signals, so the overall score is always a reproducible weighted average.
FIND pillar — discoverability (5 sub-signals)
Measures how well AI engines understand, reference, and recommend your brand.
Structured data quality — JSON-LD types, entity completeness, quantifiable attributes (prices, specs, availability).
Whether AI engines recognise and recommend your brand, based on reputation signals from ChatGPT, Claude, Gemini, Perplexity, and Grok.
2026 Scientific Standard signals: quantitative density (TR1), expert attribution (TR2), token efficiency / Flesch-Kincaid readability (TR5), answer-first structures (TR8), atomic content blocks (TR9).
How prominently the brand appears in AI engine responses — sentiment, recommendation frequency, and authority scores across all five engines.
How the brand compares to competitors named by AI engines.
USE pillar — Action Readiness (5 sub-signals)
Measures how ready the website is for autonomous AI agents.
robots.txt AI bot access, sitemap quality, llms.txt (primary signal — structured sections score highest), ai-plugin.json, .well-known/ai.json. Bonus: markdown / YAML page alternates and og.json.
AggregateRating / Review schema, third-party review platform links, evidence-of-facts.json with source-backed claims and verified_by metadata, certification schema.
Metadata completeness, Schema.org depth and entity completeness, atomic content hierarchy, compliance signals (privacy policy, GDPR, ethical certifications).
OpenAPI / GraphQL endpoint discoverability, dedicated pricing page, Product schema with populated price and availability, comparison structures (tables, pros/cons).
Sitemap quality (lastmod/priority), heading hierarchy, semantic HTML elements, canonical URLs, BreadcrumbList schema.
Action plan priorities
Each audit generates a prioritised action plan. Priority levels and their expected score impact are calibrated to be consistent:
AI providers used
SitePulsar sends audit data to the following AI providers to generate scores and recommendations:
- OpenAI — GPT-4o for reputation, competitor, and AEO aggregation
- Anthropic — Claude (Sonnet) as primary or fallback provider
- Google — Gemini 2.0 Flash for reputation and competitor analysis
- xAI — Grok for reputation analysis
- Perplexity — Brand mention and reputation queries
Only publicly available website data (crawled HTML, schema markup) and the submitted URL and brand name are sent to these providers. No personal data of your website's users is included. See our Privacy Policy for full details.