AI Insights (AEO/GEO)
Continuously monitor and improve your website's readiness to appear in AI-generated search results.
AI Insights analyzes your website's readiness for AI-driven discovery, citation, and generative search. It audits your site across 50+ factors and produces an AI Readiness Score with a prioritized action plan.
Unlike a one-time report, Evergreen monitors your AI search readiness on every crawl — so you can track regressions, improvements, and ranking shifts over time rather than only catching problems at launch.
What It Does
When you run an AI analysis, Evergreen evaluates your website across four categories:
Technical Accessibility (30% of score)
Checks whether AI systems can access and parse your content:
- llms.txt — Presence and structure of the AI summary file at your domain root
- llms-full.txt — Extended version with detailed content taxonomy
- AI crawler permissions — Whether your robots.txt allows GPTBot, ClaudeBot, PerplexityBot, OAI-SearchBot, Amazonbot, and other AI crawlers
- Page weight — Content-to-markup ratio and JavaScript dependency
- H1 structure — Heading presence and hierarchy
- Meta directives — Noindex tags and other meta-level signals
Content Engineering (30% of score)
Evaluates whether your content is structured for AI extraction and citation:
- Entity density — Proportion of proper nouns and specific terms (optimal: ~20%)
- BLUF pattern — Whether pages lead with a direct answer (Bottom Line Up Front)
- Statistics and data points — Presence of specific, citable numbers and research
- Expert quotations — Attribution signals and authorship credentials
- Readability — Flesch Reading Ease (60-70) and Gunning Fog Index (8-10)
- AI-ism detection — Flags generic filler phrases that signal low-value content
Schema & Structured Data (25% of score)
Validates machine-readable markup that connects your content to the knowledge graph:
- JSON-LD presence — Valid structured data on each page
- @id attributes — Unique entity identifiers for disambiguation
- sameAs links — Connections to Wikipedia, Wikidata, and other knowledge bases
- knowsAbout — Explicit expertise declarations in schema markup
- dateModified — Content freshness signals (60-day threshold)
- Topic relationships — about/mentions schema properties
Agent Readiness (15% of score)
Checks whether AI agents can interact with your site programmatically:
- MCP server — Model Context Protocol endpoint availability
- AI crawler configurations — Fine-grained crawler-specific rules
AI Readiness Score
Your composite score ranges from 0 to 100 and maps to a maturity level:
| Score | Level | Meaning |
|---|---|---|
| 81-100 | Leading | Optimized for AI discovery and citation |
| 61-80 | Optimized | Strong foundation, room for improvement |
| 41-60 | Structured | Basic elements in place, significant gaps |
| 21-40 | Reactive | Minimal readiness, likely invisible to AI |
| 0-20 | Unaware | No optimization, immediate action needed |
Where to Find It
Navigate to your project, then go to Insights > AI (AEO, GEO) in the sidebar.
AI Insights vs. Ask Eve
AI Insights and Ask Eve both use AI to help you understand your site, but they serve different purposes:
| AI Insights | Ask Eve | |
|---|---|---|
| What it does | Scores your site across 50+ AEO/GEO criteria | Answers freeform questions about crawl, audit, and analytics data |
| Output | Structured scored report with actionable recommendations | Conversational answers with supporting data |
| Best for | Structured AI readiness analysis and improvement planning | Quick questions, exploration, on-demand audits |
See Ask Eve for conversational AI access to your site data.
Next Steps
- Running an AI Analysis — How to start an analysis and what to expect
- Understanding Your Scores — Detailed breakdown of each category and how to improve
- AEO/GEO Guide — In-depth guide to AI search optimization strategy
- AI Tool Integrations — Query your AI Insights history from Claude, Cursor, and other AI tools
