The Stobo Score measures 7 technical factors that control whether AI systems can discover, crawl, and cite your content. These factors are AI crawler access, llms.txt, Schema markup, XML sitemap, FAQ structure, direct answers, and content freshness. Sites with all seven see citation gains within 16 days to one month.

TL;DR

  • Stobo Score measures 7 technical factors that control whether AI systems can cite your content.
  • Core components are AI crawler access, llms.txt, Schema markup, XML sitemap, FAQ structure, direct answers, and content freshness.
  • Results timeline is 16 days to one month after you fix issues. AI systems need time to recrawl your content.
  • Each factor carries weighted importance based on citation data from ChatGPT, Perplexity, and Claude.

Two Pillars of AI Engines Optimization

AEO has two pillars. We measure one of them.

Pillar 1: Onsite Optimization. Your technical foundation. Can AI crawlers access your site? Can they understand your business? Can they extract information cleanly? This determines if your content is eligible for citation.

Pillar 2: Offsite Authority. Your reputation and reach. Citations from Reddit, review platforms, and trusted publications. This determines if AI systems choose you over competitors.

We measure Pillar 1 only. We check whether ChatGPT, Claude, Perplexity, and other AI systems can access and extract your content. We don't measure if they will cite you. That depends on factors beyond technical optimization.

Building tools for Pillar 2. Stay tuned.

What Your Score Means

A high score means AI systems can find you, understand you, and cite you. Whether they actually do depends on content quality, domain authority, and competitive factors.

A score of 90 means you're technically ready. Your content is accessible. Your schema is in place. AI crawlers can extract what they need.

A score of 50 means barriers exist. AI systems struggle to access or understand your content. Technical fixes come first, then content improvements.

What We Measure

We check 7 technical elements. Each one removes a barrier to AI visibility.

1. AI Crawler Access (robots.txt)

What we check: Does your robots.txt file allow AI crawlers to access your site?

Why it matters: Block AI crawlers and you get zero citations. Content quality becomes irrelevant.

Cloudflare tracked AI crawler traffic from May 2024 to May 2025. It grew 305%. GPTBot handles 30% of all AI crawler activity alone.

OpenAI runs three separate crawlers. GPTBot collects training data. OAI-SearchBot powers search features. ChatGPT-User handles live browsing. Block the wrong one and your content disappears from AI results.

Research:

We check for 7 critical AI crawlers plus 3 recommended crawlers.

2. Machine-Readable Business Context (llms.txt)

What we check: Do you have an llms.txt file that tells AI systems about your business?

Why it matters: llms.txt gives AI a curated summary of your site formatted for machines. Jeremy Howard of Answer.AI proposed the standard in September 2024.

Major LLM providers don't officially support it yet. Early adopters report better AI understanding of their business context. Over 844,000 sites now use it. Anthropic, Cloudflare, Stripe, and Zapier all have one.

Research:

We check file structure, content quality, and resource links.

3. Structured Data (Schema Markup)

What we check: Do your pages include Schema.org markup?

Why it matters: Schema markup has the strongest link to AI citations. Stronger than any other technical factor.

Search Engine Land ran a controlled test in October 2025. Three identical pages. Different schema setups. Only the page with proper schema appeared in AI Overviews.

Microsoft's Fabrice Canel confirmed it at SMX Munich 2025: "Schema markup helps Microsoft's LLMs understand your content."

The numbers are clear. FAQPage schema gets 41% citation rates. Plain content gets 15%. That's a 2.7x advantage according to Relixir research.

Profound analyzed 680 million citations. FAQ schema pages get cited far more often. The structure removes guesswork for AI systems.

Research:

We check for Organization, Product, FAQPage, and other schema types. Having schema on multiple pages matters more than just your homepage.

4. Sitemap Configuration

What we check: Do you have a valid sitemap.xml at your domain root?

Why it matters: Sitemaps help AI crawlers find your pages faster. Less critical than other factors, but missing sitemaps hurt visibility for AI systems that can't run JavaScript.

We verify your sitemap exists, uses valid XML, and is accessible to crawlers.

5. FAQ Content Structure

What we check: Do you have FAQ content structured for AI extraction?

Why it matters: FAQ format matches how AI systems structure answers. The Princeton GEO study found 78% of AI answers use list formats. Clear headings and short paragraphs beat long narrative prose.

The business impact is measurable. GreenBananaSEO case studies show AI traffic converts 4-9x better than organic search. One financial advisor gained $4.8 million in new assets from AI referrals in 120 days. AI visitors arrive educated and ready to buy.

Research:

We check FAQ presence, question count, answer length (40-60 words is optimal), and FAQPage schema pairing.

6. Direct Answer Optimization

What we check: Does your first paragraph directly answer "What is this?"

Why it matters: AI systems focus on your first 40-60 words when picking sources. The Princeton study found direct answers get 67% more citations. This length gives enough context without getting truncated.

Research:

We check word count and clarity. Does your opening answer what users are really asking?

7. Content Freshness

What we check: Do your pages show when content was last updated?

Why it matters: AI systems prefer fresh content. Recent update dates signal trustworthy information. This matters most for pricing, features, and fast-changing topics.

How we score freshness:

  • 100 points: Updated within 30 days
  • 80 points: Updated within 90 days
  • 60 points: Updated within 180 days
  • 40 points: Updated within one year
  • 20 points: Older than one year
  • 0 points: No freshness signals found

We check HTTP headers, meta tags, Schema.org dates, and visible dates in your content.

How We Calculate Scores

We score each factor from 0-100. Then we combine them using weights based on citation impact.

The calculation includes:

  • Weighted averaging across all 7 factors
  • Bonus points when fixes work together (FAQ content paired with FAQPage schema)
  • Penalty points for critical gaps
  • Multipliers that reduce scores when basic access issues exist

Your Grade

90-100 (Grade A): Your technical foundation is solid. AI crawlers can access and extract your content. Focus on content quality and building authority.

80-89 (Grade B): Good foundation with minor gaps. Fix the flagged issues to reach Grade A. Most sites need 2-4 hours of work.

65-79 (Grade C): Multiple improvements needed. Plan for 8-16 hours to fix core gaps.

50-64 (Grade D): Serious technical barriers. AI systems struggle to access your content. Fix critical issues first.

0-49 (Grade E): Critical failures block AI visibility. Content improvements won't help until you fix these basics.

What We Don't Measure

The Stobo Score evaluates technical foundation only. We don't measure:

  • Content quality or depth
  • Domain authority or backlinks
  • Brand recognition
  • Review platform presence
  • Wikipedia listings
  • Actual citation frequency

These factors affect AI citations. We can't detect or fix them automatically. Many fall under Pillar 2: Offsite Authority. We're building tools to help with that.

Validation

We improve our methodology based on data. As we learn how technical scores connect to real citations, we update weights and rules.

Our scoring system is versioned. We document all changes. Historical scores get recalculated so you can compare over time.

Current version (91.0, December 2025) includes:

  • Research-backed weights for each factor
  • FAQ-schema pairing based on the Relixir 2.7x study
  • Page-level schema checks (not just homepage)
  • Updated crawler lists for the current AI landscape

Research Foundation

Our methodology builds on peer-reviewed research:

  1. Princeton/Georgia Tech GEO Study (ACM SIGKDD 2024). Tested 10,000 queries.

  2. Cloudflare AI Crawler Analysis (2024-2025). Tracked crawler growth patterns.

  3. Profound Citation Research (680M citations). Proved FAQ schema advantage.

  4. Search Engine Land Schema Experiments (2024-2025). Showed schema impact on AI Overviews.

  5. Conductor Benchmarks (13,770 domains, 3.3B sessions). Measured conversion rates.

  6. GreenBananaSEO Case Studies. Real ROI data.

We update methodology as new research emerges.

Related Guides

Questions

Questions about your score? Contact us at hello@trystobo.com. We respond within 24 hours.