GEO & AI-Discovery Audit

1. Executive Summary

The content foundation here is genuinely strong for GEO: 11 sector pages each carry evidence-dense FAQs, 9 gated Executive Reports cite named, dated sources, and the blog already functions as a topic hub. The gap is almost entirely technical and structural, not creative — the site is not yet built to be read by a crawler or an agent the way it is by a human. Two findings below are P0 because they affect every single page at once.

2. What the Current Research Actually Says

GEO/AEO is a genuinely new field with a lot of agency marketing noise around it. Filtering that out, five points hold up across Google's own documentation, Microsoft's public statements, and the independent research (including the Princeton/Georgia Tech/IIT Delhi study that coined the term "GEO"):

1. It is not a separate ruleset. Google's own AI-optimization guide says its generative features "are rooted in our core Search ranking and quality systems" — there is no secret AI checklist, and the single test Google recommends is "would a visitor find this satisfying?"
2. E-E-A-T is the filter, not keywords. Both Google's AI Overviews and independent audits of ChatGPT/Perplexity citation behavior consistently point to Experience, Expertise, Authoritativeness, Trust — author identity, named sources, and verifiable specifics beat keyword density.
3. Structure earns citations, but isn't required. The GEO research found definition-first sentences, named statistics, and question-phrased headings lift citation rates 30–115%; a Search/Atlas 2024 study separately found no correlation between schema volume alone and citation rate. Read together: structure content to be citable; treat schema as hygiene that removes friction, not a lever that creates citations by itself.
4. Schema is confirmed to matter for machine understanding, if not ranking. Microsoft's Bing lead and Google's structured-data engineer both confirmed on the record in 2025 that their systems use schema to understand pages; one 2026 dataset found ~71% of pages ChatGPT cites, and ~65% of pages Google AI Mode cites, carry structured data. Correlation, not proof of causation — still worth doing.
5. Crawlability is the precondition everyone skips. Many sites unintentionally block AI crawlers via robots.txt or a CDN default (Cloudflare's own default now blocks AI bots). llms.txt is an emerging, low-cost, voluntary standard (not a Google ranking factor, but already honored by GPTBot, ClaudeBot, Google-Extended, and PerplexityBot) that points agents at a site's most important pages in one clean Markdown file.

3. Site Audit — Current State

Content architecture (strength)

42 pages organized into a real hub-and-spoke structure: a University homepage, an OS platform page, 3 program pages, 11 industry pages (each with 12 numbered sections including FAQs), 9 gated Executive Reports, a Blog acting as a resource hub, Case Studies, and a Community layer (Members / Initiatives / Events). This is a legitimately strong starting shape for topical authority — most sites this audit sees have to build it from nothing.

Technical gaps (the P0 list)

4. Prioritized Recommendations

5. Revised Sitemap & Content Architecture

Current structure is sound; the additions below close the topical-authority gaps a buyer's AI research session would otherwise hit a wall on.

New pages recommended:

• Standalone About / Company page — right now "About" is only an anchor on the homepage. A dedicated page with the brand-entity statement, founder bio, and Organization schema gives AI systems one authoritative place to answer "who is World AI X."
• Glossary hub — one pillar page per core term (Chief AI Officer, AI operating model, AI-native leadership, AI transformation framework), each a self-contained, citable definition.
• 2–3 comparison pages — the format AEO research shows AI systems reach for most: "CAIO vs. CTO vs. CIO," "AI operating model vs. AI strategy consulting," "in-house AI training vs. certification."
• FAQ hub aggregating the strongest Q&As already written across the 11 industry pages, cross-linked back to each source page.

6. Page-Level Playbook

Full breakdowns for the seven page archetypes below; the matrix at the end of this section applies the same <title>/description pattern to every actual page.

Title / meta-description matrix (apply at deploy)

7. Entity & Brand Authority

The single highest-leverage fix here is narrative, not technical: publish one sentence, reused everywhere, describing how World AI X, World AI University, the AI Transformation OS, and the World AI Council relate — homepage, footer, the new About page, and inside the Organization schema's description field. Consistency is the signal; AI systems trust brands whose name, description, and relationships read identically wherever they appear.

• Add a Person schema + short bio for the founder/instructor identity behind the program (the booking links already point to a named individual) — author and instructor identity is a direct E-E-A-T signal.
• Keep program names (CAIO Accelerator, AI Transformation OS, etc.) capitalized and worded identically across every page, the schema, and any off-site profile (LinkedIn, PR).
• List every named client/member logo consistently — these double as third-party entity corroboration once those organizations' own sites/press mention the relationship.

8. Technical SEO, Schema & AI-Agent Requirements

Head tags: this project's current runtime generates a shared, minimal <head> (charset + viewport only) for every page — verify with your dev team whether per-page <title>, meta description, canonical, Open Graph, and JSON-LD can be injected at the template layer; if not, add a thin server-side (or static-export) layer that injects these per route before the page ships to production. This is likely the single highest-ROI engineering task on this whole list.

robots.txt, sitemap.xml, llms.txt — drafted and added to this project (see the project's root files); deploy all three at the real domain root once it's finalized, and update the placeholder waiu.org domain used in them to the true production host.

Ready-to-paste JSON-LD (fill in the bracketed values):

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "World AI X",
  "url": "https://u.worldaix.com/",
  "logo": "https://u.worldaix.com/assets/worldaix-black.png",
  "description": "World AI X certifies AI-native leaders through World AI University and runs the AI Transformation OS, the governed platform for enterprise AI initiative portfolios.",
  "sameAs": ["[LinkedIn URL]", "[X/Twitter URL]", "[Wikidata URL once created]"]
}
</script>

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Course",
  "name": "CAIO Accelerator",
  "description": "6-week certification program for executives becoming AI-native leaders.",
  "provider": { "@type": "Organization", "name": "World AI University", "sameAs": "https://u.worldaix.com/" },
  "hasCourseInstance": { "@type": "CourseInstance", "courseMode": "blended", "startDate": "2026-09-21" }
}
</script>

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    { "@type": "Question", "name": "[existing FAQ question text]",
      "acceptedAnswer": { "@type": "Answer", "text": "[existing FAQ answer text]" } }
  ]
}
</script>

AI-agent task readiness: the site already has real, working action affordances an agent can follow — an Apply form, a Login link, and a book-a-call link. Keep these as plain, crawlable <a href> links (not JS-only onClick handlers with no href) so an agent parsing raw HTML can find and follow them without executing JavaScript.

9. Evidence & Citation Readiness

• The Executive Reports already cite named, dated sources (Doximity, McKinsey, IEA, GAO, etc.) in a numbered reference list — this is exactly the format research shows AI systems prefer. The one gap: none of those citations link out to a live source URL. Add the hyperlink to each.
• Add a visible "last reviewed/updated" date to every Report and Blog post — freshness is a repeatedly-cited factor in maintaining AI citation priority over time.
• Audit Case Studies for any anonymized ("a Fortune 500 healthcare system") entries and replace with named clients wherever permission allows — named, specific proof outperforms generic proof in every E-E-A-T-driven system.

10. Content Gaps & New Opportunities

11. Off-Site Discoverability

• Create a Wikidata entity for World AI X / World AI University once the entity statement (§7) is finalized — Wikidata is a primary source many LLMs are grounded on.
• Keep a consistent, active LinkedIn presence for both the company and the named founder/instructor — publish Executive Report findings there to earn independent citations pointing back.
• Pursue listings in the relevant sector associations/directories for each of the 11 industries (e.g., health-IT, GovTech, EdTech directories) — third-party mentions are what "entity stacking" research says does the real heavy lifting for AI trust, more than schema alone.
• Pitch the Executive Reports' original data points to trade press and podcasts as syndicated research — earned coverage citing "World AI X research" is exactly the kind of external corroboration that compounds AI-citation authority over time.

12. Measurement Framework

13. 90-Day Roadmap

Sources

Google Search Central, Guide to Optimizing for Generative AI Features (developers.google.com) · Search Engine Land, AI Overviews optimization guide and How schema markup fits into AI search — without the hype · Semrush, What Is LLMs.txt & Should You Use It? · derivatex.agency, How Schema Markup Affects LLM Citation (Fabrice Canel/Microsoft, Ryan Levering/Google statements; SEranking citation-rate dataset) · red-engage.com GEO guide (Princeton/Georgia Tech/IIT Delhi GEO research, 30–115% visibility finding) · Frase and HubSpot AEO guides · Amsive, Answer Engine Optimization (brand AI-visibility share data) · llmrefs.com and Firecrawl, llms.txt implementation guides.