L0 → L1
Phase 1 — Quick Wins
Fix crawler access, robots.txt, render gaps, and meta signals. No code required. Done in a week.
GEO Readiness Playbook
From AI-invisible
to fully transactable.
The step-by-step framework to make your site discoverable, citable, and actionable by AI systems. Four phases. Five readiness levels. Every step linked to the audit, checklist, and methodology.
Harshak Patel runs Hidden Layer, where the work is auditing how AI systems surface — or refuse to surface — brands and products. Every phase is grounded in the same 46-check audit Hidden Layer runs as its product. Full profile →
“The shift from search to AI discovery is as significant as the shift from print to digital — brands that don't adapt will become invisible.”
a16z, 2026 AI Commerce Report
“We're seeing a fundamental change in how people find information. Generative AI isn't just a feature — it's the new interface layer.”
Google, May 2026
The readiness ladder
Five sequential levels. Each depends on the one below. Most sites are stuck at L0 — blocked before they're even seen.
TL;DR
Fix access first (Phase 1), then signal quality (Phase 2), then content authority (Phase 3), then agentic integration (Phase 4). Skipping levels doesn't work — an uncrawlable site cannot be cited regardless of content quality.
- FL0 — InvisibleAI crawlers cannot read your site. JavaScript-only content, blocked bots, missing server-render.
- D–CL1 — DiscoverableCrawlers can access you. robots.txt declared. Server-rendered HTML. Basic meta signals present.
- C–BL2 — IncludedStructured data complete. Merchant feeds submitted. Products eligible for AI shopping surfaces.
- B–AL3 — CitedBrand earns LLM mentions. Entity present in training data. Prose answers real buyer questions.
- AL4 — TransactableAgents can discover, recommend, and complete transactions with your catalogue in real time.
How AI finds and cites you
Every AI citation follows the same pipeline. GEO is about optimising each stage.
User querynatural language
AI systemChatGPT / Perplexity…
RetrievalRAG / web search
Your pageserver-rendered HTML
Rankedrelevance scoring
Synthesisanswer generation
Cited ✦
1
Crawl access
robots.txt, llms.txt, server-rendered HTML — Phase 1 fixes these.
2
Retrieval ranking
Structured data (schema.org), feeds, and entity links — Phase 2.
3
Citation selection
Statistics, citations, and authoritative prose — Phase 3.
How GEO differs from SEO and AEO
Same web, three optimisation targets. GEO is about being cited inside the answer — not ranking beside it.
| SEO | AEO | GEO | |
|---|---|---|---|
| Optimises for | Ranking in the blue-link results | Being the direct answer | Being cited inside the AI-generated answer |
| Primary surface | Google / Bing results page | Featured snippets, People-Also-Ask, voice | ChatGPT, Perplexity, Gemini, AI Overviews |
| Consumed by | A human scanning links | A human wanting a quick answer | An LLM synthesising a response |
| Unit of success | Position & organic clicks | Owning the answer box | Citation / mention rate (often zero-click) |
| Key levers | Keywords, backlinks, page speed | Structured Q&A, concise answers, schema | Structured data, entity presence, cited stats, crawlability |
The four phases
Work the phases in order. Each phase unlocks the next readiness level.
L1 → L2
Phase 2 — Schema & Feeds
Organization schema, FAQPage, Product schema, Merchant Center, RSS. Makes products AI-eligible.
L2 → L3
Phase 3 — Content & Citation
Earn LLM recommendations with statistics-rich prose, entity presence, and AI referral tracking.
L3 → L4
Phase 4 — Agentic & Advanced
WebMCP annotations, /.well-known/webmcp manifest, x402 / BuyAction for agent-initiated commerce.
Know if it's actually working
The work only counts if you can see it move. Measure with the same audit the playbook is built around — and avoid the traps that make GEO scores lie.
Re-run your auditYour measurement instrument — the same AI-readiness score the playbook targets. Re-run after each phase to see the delta.Track unbranded mentions61.7% of AI citations never name the brand. Count mentions, not just links, or you will under-measure your wins.Don't trust a single runOne prompt, one model, one day is noise. Probe repeatedly — and expect 12–24 months for training-data presence to move.
Go deeper
Glossary13 GEO terms defined — from readiness levels to WebMCP and cold recall.Print editionComplete playbook as a single print-optimised page — download as PDF.ChecklistStep-by-step checklist for every audit dimension.MethodologyHow Hidden Layer scores AI readiness.BlogDeep-dives on GEO tactics and research.Run an auditGet your domain's AI readiness score instantly.