Skip to content
Guides/AI Readiness Score
methodologyNew

AI Readiness Score: What It Measures and How to Improve It

A complete breakdown of all 6 scoring categories, every check, grade thresholds, and exactly what to fix to move the needle — based on the actual Open Shadow scanner methodology.

📊 10 min read·Methodology v2.0·March 2026

The Score at a Glance

The AI Readiness Score is a 0–100 metric across 6 categories and 20 individual checks. Points are native — they sum directly to your score with no normalization. Every check cites a published standard or documented AI crawler behavior.

Score Distribution
🔍
AI Discoverability25 pts
⚙️
Technical Foundation20 pts
📄
Content Structure20 pts
📈
SEO Fundamentals15 pts
🛡️
Security Posture10 pts
Accessibility10 pts

Grade Thresholds

A
90–100
B
75–89
C
50–74
D
30–49
F
0–29

Category Breakdown

🔍

1. AI Discoverability/ 25 pts

This is Open Shadow's differentiator. Measures whether AI systems — ChatGPT, Perplexity, Claude — can find, crawl, understand, and correctly attribute your content. Most SEO tools don't score this at all.

AI bots can access your site5 pts

In your robots.txt, allow at least 4 of the major AI crawlers (GPTBot, ClaudeBot, Google-Extended, PerplexityBot, Applebot-Extended). Score: 5pts if 4+ are allowed, 3pts if 2–3 are allowed, 2pts if robots.txt exists but bot rules are unclear, 0pts if all bots are blocked.

Has llms.txt for AI systems5 pts

Create a /llms.txt file at your domain root. It tells AI assistants what your site is about, which pages matter, and what to ignore. Use the Open Shadow llms.txt Generator to build one in 60 seconds.

XML sitemap quality4 pts

Have an XML sitemap at /sitemap.xml and submit it to Google Search Console. Score: 4pts for a valid sitemap with 5–50,000 URLs, 2pts for sitemap exists but outside range, 0pts for missing.

Rich structured data (JSON-LD)4 pts

Add JSON-LD schema markup to your pages. Score: 4pts for 3+ schema types, 2pts for 1–2 types, 0pts for none. Recommended schemas: Organization, WebSite, WebPage, Article, FAQPage, BreadcrumbList.

Clean content for AI extraction3 pts

Use a single H1, proper heading hierarchy (H1→H2→H3), and write at least 300 words of meaningful content per page. Score: 3pts for both, 1pt for one condition, 0pts for neither.

Metadata for AI attribution2 pts

Every page needs a meta description and a canonical URL. These are how AI systems attribute and cite your content. Both present = 2pts, one present = 1pt, neither = 0pts.

Explicitly manages AI bots2 pts

Your robots.txt should explicitly name at least one AI bot (to allow or block). "Unmanaged" access — where you've never thought about which bots can crawl your site — scores 0pts here. Even blocking GPTBot scores 2pts: it shows intentionality.

⚙️

2. Technical Foundation/ 20 pts

AI crawlers and search bots prioritize technically sound sites. A site that's slow, HTTP-only, or renders incorrectly on mobile gets deprioritized in crawl queues.

HTTPS enabled8 pts

This is the single biggest check in the entire scorer (8pts). Serve your site over HTTPS. Almost all hosting providers offer free SSL via Let's Encrypt. There is no partial credit — HTTP sites score 0 here.

Good performance score6 pts

Score measured via Lighthouse-equivalent analysis. 6pts for score ≥80, 3pts for score ≥50 or fast initial response. Improve by: compressing images (WebP), minifying CSS/JS, using a CDN, and deferring non-critical scripts.

Mobile-friendly viewport3 pts

Add <meta name="viewport" content="width=device-width, initial-scale=1"> to your HTML <head>. Google uses mobile-first indexing, meaning the mobile version of your page is what's indexed and scored.

Valid SSL certificate3 pts

Your SSL certificate must be valid, not expired, and correctly chained. Check it at ssllabs.com/ssltest. Let's Encrypt certificates auto-renew if configured correctly — add monitoring so you're not caught out.

📄

3. Content Structure/ 20 pts

AI systems parse HTML to extract meaning. Pages with clean structure, semantic markup, and sufficient content depth are dramatically more likely to appear in AI-generated answers.

Has JSON-LD structured data5 pts

Add at least one JSON-LD <script type="application/ld+json"> block to your pages. The minimum useful schema is Organization or WebSite. This check overlaps with AI Discoverability intentionally — it's that important.

Multiple schema types defined5 pts

3+ schema types = 5pts, 1–2 types = 2pts, none = 0pts. Beyond Organization and WebSite, add: Article or BlogPosting for content pages, FAQPage for FAQ sections, BreadcrumbList for navigation, Product for e-commerce, and LocalBusiness for physical locations.

Proper heading hierarchy4 pts

Each page should have exactly one H1 tag, followed by H2s for major sections, H3s for subsections. Never skip levels (don't go H1→H3). AI systems use heading structure as a content outline when deciding what to include in answers.

Images have alt text (>90%)3 pts

90%+ of images with alt text = 3pts, 50–90% = 2pts, below 50% = 0pts. Alt text isn't just an accessibility requirement — it's how AI systems understand your visual content. For decorative images, use alt="".

Sufficient content depth3 pts

500+ words = 3pts, 200–500 words = 1pt, below 200 = 0pts. Thin pages rarely appear in AI-generated answers. AI systems favor pages that provide comprehensive answers, not thin landing pages.

📈

4. SEO Fundamentals/ 15 pts

AI search builds on traditional SEO signals. A page that ranks well in Google is more likely to be surfaced by AI systems — but several SEO elements (descriptions, canonical URLs) have direct value for AI citation independent of traditional rankings.

Optimized title tag5 pts

5pts for present and 50–60 characters, 3pts for present but outside ideal length. Your title tag is the first thing AI systems use to understand a page's topic. Keep it specific, descriptive, and front-load the most important keywords.

Optimized meta description5 pts

5pts for present and 150–160 characters, 3pts for present but outside ideal length. Meta descriptions are often used verbatim by AI systems as page summaries when citing your content.

Has canonical URL3 pts

Add <link rel="canonical" href="https://yourdomain.com/page"> to every page. This tells crawlers which version of a URL is authoritative, preventing duplicate content confusion when AI systems index your site.

Complete Open Graph tags2 pts

Add og:title, og:description, og:image, and og:url. Social platforms and AI systems use Open Graph data to generate rich previews and understand page context.

🛡️

5. Security Posture/ 10 pts

Security headers are low-effort, high-signal. Sites with proper security configuration are treated as more trustworthy by crawlers. Missing headers like CSP or HSTS are a red flag for AI systems assessing site quality.

Security headers configured6 pts

6pts for a security score ≥70, 3pts for ≥40, 0pts below 40. Key headers to add: Content-Security-Policy, X-Frame-Options, X-Content-Type-Options, Referrer-Policy, and Permissions-Policy. Check your current headers at securityheaders.com.

HSTS with strong max-age4 pts

4pts for HSTS max-age ≥31536000 (1 year), 2pts for HSTS present but shorter. Add to your server config: Strict-Transport-Security: max-age=31536000; includeSubDomains; preload. On Vercel, set this in vercel.json headers.

6. Accessibility/ 10 pts

Accessible HTML uses semantic elements that AI parsers read more accurately. The lang attribute helps language detection; ARIA landmarks provide structural context; high accessibility scores correlate with clean, well-structured HTML that AI systems handle better.

Language attribute set3 pts

Add lang="en" (or your language code) to the <html> element. This is a WCAG 3.1.1 requirement and also helps AI systems determine content language for multilingual indexing.

ARIA landmarks present3 pts

Use semantic HTML elements with built-in ARIA landmark roles: <header>, <nav>, <main>, <footer>, <aside>, <section>. These tell AI parsers exactly which part of the page is content vs. navigation vs. supplementary information.

Good accessibility score4 pts

4pts for score ≥80, 2pts for ≥50. Check with axe-core, WAVE, or Lighthouse accessibility audit. High accessibility scores virtually always correlate with clean, semantic HTML — the exact structure AI systems parse most accurately.

Fastest Path to a Higher Score

If you're starting from a D or C, here's the order that gets you the most points for the least effort:

+8
~10 min

Enable HTTPS

If you're still on HTTP, this is your first move. Nearly every host offers it free.

+5
~5 min

Add llms.txt

Use the Open Shadow llms.txt Generator. Drop the file in your public/ root. Done.

+5
~15 min

Allow AI crawlers in robots.txt

Add explicit User-agent rules for GPTBot, ClaudeBot, Google-Extended, PerplexityBot.

+5
~20 min

Add JSON-LD schema

Start with Organization + WebSite. Use schema.org to build the JSON, paste into a <script> tag.

+5
~10 min

Write a proper title tag

50–60 characters, descriptive, front-loaded. One change, 5 points.

+5
~10 min

Write a proper meta description

150–160 characters. This doubles as what AI systems show when citing your page.

+3
~5 min

Add a canonical URL

<link rel="canonical" href="..."> in every page's <head>.

+3
~5 min

Add lang="en" to <html>

One attribute. 3 points. Set it and forget it.

+4
~30 min

Fix security headers

Add X-Frame-Options, X-Content-Type-Options, Referrer-Policy. HSTS adds another 4pts.

Frequently Asked Questions

What is the AI Readiness Score?+

A 0–100 metric across 6 categories and 20 checks that measures how well your site is configured for AI crawlers, AI search, and AI-generated answers. Run it free at openshadow.io/check.

What's the single biggest point gain available?+

HTTPS — worth 8 points, the highest single check. If you're serving over HTTP, that's 8 points you're leaving on the table for what amounts to a 10-minute fix.

Does a better score mean better Google rankings?+

Not directly, but there's significant overlap. Many checks (HTTPS, title tags, meta descriptions, sitemap, canonical URLs) are also traditional SEO signals. Improving your AI Readiness Score often improves traditional SEO simultaneously.

What's AI Discoverability and why does it matter most?+

AI Discoverability (25pts) measures whether systems like ChatGPT, Perplexity, and Claude can find and accurately represent your site. It's unique to Open Shadow — no traditional SEO tool scores this. A site invisible to AI systems misses the fastest-growing traffic source in 2026.

How often should I re-scan?+

After any major deployment or content change. Scan results are cached for 1 hour, so wait at least that long after making changes before re-scanning to see updated scores.

My score is an A. What now?+

Check the AI Perception module — it tests how ChatGPT, Claude, and Perplexity actually describe your brand right now. Technical readiness and AI perception are related but different: a site can score an A and still be described inaccurately by AI systems.