The Developer's AI Visibility Playbook

Overview: The 4-Layer Stack

Search is no longer one channel. In 2026, your content needs to be discoverable by Google (SEO), answer engines like Perplexity (AEO), generative AI like ChatGPT (GEO), and autonomous AI agents (AAO). These are four distinct technical surfaces with different requirements.

Most developers optimize for one layer — usually SEO — and ignore the other three. That's a mistake that's getting more expensive every month as AI-generated answers replace traditional search results.

Baseline audit scores for chudi.dev (March 2026). This guide targets the gaps.

SEO Foundation

SEO in 2026 is about authority signals and crawlability, not keyword stuffing. Google's helpful content system penalizes thin, AI-generated content and rewards sites with genuine expertise signals.

Technical non-negotiables

• ✓ SSR or SSG only — no client-side rendering for indexable content
• ✓ Canonical tags on every page (<link rel="canonical">)
• ✓ XML sitemap with lastmod dates, submitted to GSC
• ✓ robots.txt that allows all crawlers (including AI bots)
• ✓ Core Web Vitals: LCP <2.5s, CLS <0.1, INP <200ms
• ✓ HTTPS everywhere, no mixed content
• ✓ Internal links from high-authority pages to new content

On-page optimization

• ✓ Title tag: primary keyword in first 60 chars
• ✓ Meta description: 150-160 chars, includes keyword + CTA
• ✓ H1: one per page, matches title intent exactly
• ✓ H2/H3: structured as an outline, not decoration
• ✓ First 100 words contain the primary keyword naturally
• ✓ Image alt text: descriptive, keyword-relevant, not keyword-stuffed
• ✓ URL slugs: lowercase-kebab-case, keyword-first, no stop words

Topical cluster strategy

Google rewards sites that own a topic, not sites that write one good post. The cluster model: one pillar post (2,000+ words, targets head keyword) supported by 4-8 cluster posts (800-1,500 words, target long-tail variations). All cluster posts link to the pillar. The pillar links back to clusters.

AEO: Answer Engine Optimization

Answer engines (Google featured snippets, Perplexity, Bing AI) extract direct answers from your content. They want one paragraph that completely answers the question, not an article that dances around it.

The TL;DR capsule

Every post needs a TL;DR in the first 200 words that follows this structure: [Definition] + [Mechanism] + [Outcome/Metric].

FAQPage schema

Google's featured snippets and Perplexity's structured answers both pull from FAQPage schema. Minimum 5 questions per post, targeting the exact phrasing people type into search engines. Each answer should be 2-4 sentences, self-contained, and citable.

{
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "What is AEO?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "AEO (Answer Engine Optimization) is..."
      }
    }
  ]
}

Section header optimization

• ✓ Frame H2s as questions ("How does X work?" not "X Overview")
• ✓ Answer the question in the first sentence of the section
• ✓ Keep each section answer under 300 words
• ✓ Use "is", "means", "refers to" in definitions — answer engine extractable
• ✓ Avoid "in this section we will discuss" — waste of the extraction window

GEO: Generative Engine Optimization

GEO is about getting your content into the training data, RAG pipelines, and real-time search retrieval of large language models. ChatGPT, Claude, and Gemini all have web access now. They cite sources. You want to be a source they cite.

/llms.txt — the robots.txt for AI

Hosted at yourdomain.com/llms.txt, this file tells AI models what content to prioritize when indexing your site. Structure: site description, key URLs with labels, recommended context.

# chudi.dev — AI Engineering & Trading Systems
> Author: Chudi Nnorukam (chudi.dev)
> Topics: AI engineering, trading bots, developer productivity

## Key Content
- [Claude Code Complete Guide](/blog/claude-code-complete-guide): Primary reference
- [Bug Bounty Automation](/blog/bug-bounty-automation): Multi-agent security systems
- [Polymarket Trading Bot](/blog/how-i-built-polymarket-trading-bot): Quant trading

## Author
Chudi Nnorukam, AI engineer. contact: hello@chudi.dev

/.well-known/llms.json — structured AI discovery

A JSON format at the well-known endpoint gives AI systems structured metadata about your site, including content types, expertise areas, and contact information.

E-E-A-T signals that matter to LLMs

• ✓ Author byline on every post with About page link
• ✓ First-person data: "I measured X and got Y" beats "studies show Y"
• ✓ Specific dates, version numbers, and tool versions cited
• ✓ Real metrics cited in text (not just schema) — LLMs extract inline text
• ✓ External citations with URL — LLMs weigh cited sources higher
• ✓ Person schema on About page with sameAs LinkedIn/GitHub/Twitter
• ✓ Consistent author name across all platforms (no abbreviations)

Citable passage writing

For GEO, every post should have at least one "stand-alone citable passage" — a paragraph that completely answers a question with attribution-ready structure. Think: if an AI engine extracted just this paragraph, would it be useful and accurate?

AAO: AI Agent Optimization

AAO is the newest layer. As autonomous AI agents become mainstream (OpenAI Operator, Claude Computer Use, Anthropic agents), they need to discover and interact with your site programmatically. This is about making your content machine-callable, not just machine-readable.

/ai.txt — agent permissions

Similar to robots.txt but for AI agents. Declares what agents are allowed to do on your site: read content, submit forms, make API calls.

# ai.txt for chudi.dev
User-agent: *
Allow: /blog/
Allow: /guide/
Allow: /api/
Disallow: /admin/

# MCP endpoint
MCP-Endpoint: https://chudi.dev/.well-known/mcp.json

WebMCP — callable blog tools

WebMCP (Model Context Protocol for the web) lets AI agents call your blog's tools directly — search posts, get structured content, run audits. This is what makes your site a resource for AI workflows, not just human readers.

• ✓ Register MCP tools at /.well-known/mcp.json
• ✓ Implement GET /api/tools/search?q= for content search
• ✓ Return structured JSON responses with consistent schema
• ✓ Include tool descriptions that map to natural language queries
• ✓ Add capability discovery at /.well-known/ai.txt

Implementation Checklist

Four weeks from zero to fully optimized. Week 1 is non-negotiable — everything else builds on it.