Skip to main content

Articles Technical GEO

The 12-Layer Technical GEO Stack & 7-Step AI Search Funnel from Discover to Mention: Why You Must Audit Bottom-Up

Many teams rush to publish llms.txt or rewrite content while their WAF silently blocks OAI-SearchBot or critical pages rely on client-side JS. Here is why bottom-up engineering is mandatory.

When a brand is missing from ChatGPT, Perplexity, or Google AI Overview answers, the knee-jerk reaction is often to publish more blog posts or drop in an llms.txt file. Yet across dozens of international technical audits, over 70% of visibility failures stem from lower-level infrastructure layers.

To earn an AI citation, your content must pass 7 sequential gates: Discover → Crawl → Index → Retrieve → Ground → Cite → Mention. Failing at any early gate renders every downstream effort useless.

1. The 7 Gates of Generative Search Engines

Unlike classic search engines that simply rank ten blue links, generative engines execute multi-stage Retrieval-Augmented Generation (RAG) pipelines:

  • Discover & Crawl: The system discovers URLs via sitemaps, RSS, IndexNow, or links, then dispatches crawlers (OAI-SearchBot, ClaudeBot, Googlebot) to fetch the raw response.
  • Index & Retrieve: Clean text is parsed, chunked, and stored in hybrid lexical-semantic indices to be retrieved when a user prompt triggers a search.
  • Ground, Cite & Mention: The model cross-references facts across trusted sources. Pages with clean Semantic HTML, verifiable data, and unambiguous Entity signals are selected for grounding, linked as citations, and named in the final response.

2. The 12-Layer Technical GEO Stack

In the Brian Ho Marketing Academy methodology (taught in depth in our Master GEO Course), a domain's technical surface is organized into 4 tiers and 12 distinct layers:

Layer Core Technical Scope Tier
L1 - L2 DNS, Hosting, TLS/HTTPS, HTTP/3, TTFB & Server Response Latency Infrastructure
L3 CDN, WAF & Bot Management (preventing false-positive blocks on AI search bots) Infrastructure
L4 - L7 HTTP Status, Redirects, robots.txt, Content Signals, Canonicals, Internal Links, Sitemap, IndexNow Access & Discovery
L8 - L10 SSR/SSG Rendering (content in raw HTML), Semantic HTML, Snippet Controls, Schema.org & Entities Comprehension
L11 - L12 llms.txt, Agent-ready Markdown, MCP, WebMCP & Web Bot Auth Agentic

3. The Golden Rule: Audit and Fix Bottom-Up

If Layer 3 (WAF) serves a 403 or JS challenge to OAI-SearchBot, perfection at Layer 10 (Schema) or Layer 11 (llms.txt) will not save you. Similarly, if Layer 8 relies on pure Client-Side Rendering (CSR) returning an empty <div id="root"></div> shell, most AI crawlers will index nothing.

This is why the technical-geo-audit orchestrator inside Brian Ho Marketing MCP (BHM-MCP) at mcp.brianhomarketing.com enforces strict bottom-up verification with real HTTP evidence before touching editorial layers.