SEO · AEO · GEO — driven by Claude

Get your site cited by the answer, not buried under it.

AI search doesn't return 10 blue links — it writes one answer and names a few sources. This is the front-to-back method for using Claude to make your site one of the sources it names.

New here? Paste a page into the free checker for an instant AI-readiness score, or read the method below.

The short version

Point Claude at your site to do five jobs in order: audit what's blocking you, rewrite pages so answer engines can extract them, generate the technical files (AI-crawler robots.txt, llms.txt, JSON-LD), build consensus across the places AI reads, and measure whether you're getting cited. Each phase below is a page with copy-paste Claude prompts you can run today.

Why this changed

Search is now three overlapping jobs, not one. Classic SEO still wins the clicks that exist. GEO (Generative Engine Optimization) gets you cited inside AI answers. AEO (Answer Engine Optimization) structures a page so it becomes the single extracted answer. You optimize all three at once — and Google's first official generative-AI guide (May 15, 2026) says optimizing for its AI features is just SEO — no special files, no AI markup, no separate schema — a useful antidote to vendor hype.

SEO

Rank in the list of links. Still the foundation and the volume.

clicks

GEO

Get cited inside generated answers across ChatGPT, Perplexity, AI Overviews, Gemini, Claude.

citations

AEO

Be the one extracted answer — the featured snippet's AI descendant.

the answer

The numbers that define the moment (directional, sourced mid-2026, shelf life of months — re-verify before betting a budget):

The one model everything hangs on

AI answer engines look for agreement across independent sources before they cite you. If the same true facts about you appear on your site, on Reddit, in a "best of" listicle, in a review, and in a YouTube video, the model gets confident and names you. If you exist only on your own domain, it treats your claims skeptically and recommends someone with a broader footprint. Everything in the five phases serves that one idea. It's research-backed, too: a 2025 University of Toronto study found AI search overwhelmingly favors earned, third-party sources over brand-owned content, and the founding Princeton GEO study measured visibility lifts of up to 40% from adding statistics, quotations, and cited sources.

The workflow, in five phases

Run them in order. Each links to a full page with the exact Claude prompts.

0

Audit what's blocking you

Have Claude check AI-crawler access, indexing health, schema coverage, orphan pages, and stray noindex — before you write a word.

Run the audit →
1

Rewrite pages for extraction (AEO)

Front-load a self-contained, stat-bearing answer; add fact density and information gain; restructure into question headers and comparison tables.

Fix the pages →
2

Generate the technical files

AI-crawler robots.txt, a real llms.txt, accurate JSON-LD (Organization, FAQ, HowTo, Speakable), and freshness signals.

Ship the files →
3

Build off-page consensus

Put the same true story where AI reads — Reddit, listicles, reviews, YouTube — and enforce one consistent entity fact sheet everywhere.

Build consensus →
4

Measure whether you're cited

Baseline share-of-voice across engines, track citation churn (expect 40–60%), and catch AI referrals that hide in Direct traffic.

Measure it →

Why use Claude for this specifically

Because every phase is a task Claude does well against your actual files, not in the abstract. It can read a page and rewrite the first 200 words into a self-contained answer; generate valid JSON-LD and an llms.txt from your sitemap; diff your robots.txt against the current crawler list; draft an entity fact sheet you reuse everywhere; and score a batch of AI answers for whether your brand shows up. Keep your brand facts in a Project so every prompt starts from the same source of truth, and use Claude Code when you want it editing the site files directly.

Free, and the whole method

Everything here is free and complete: the full five-phase method, the 3,307-question dataset with a downloadable CSV, a working checker, and every prompt — no signup, no email wall, no "book a demo." That is deliberate. Much of the paid GEO market sells llms.txt packages, schema audits, and guaranteed-citation retainers — and this site's own guides show, with evidence, that those are not the levers that get you cited. The actual method is learnable, so we wrote it down and gave it away. If you are weighing whether to do this yourself or pay someone, we made an honest map of exactly when paying is worth it: DIY vs a paid GEO service.

Guides & references

Deeper dives on the terms, the platform differences, and the tooling — the questions people search most.

The AI-search question dataset

3,307 real questions, harvested & counted

Publish original data

Become the source AI cites — autocomplete harvesting & more

Trends, SERP APIs & GSC

Programmatic demand & citation data sources

Build topical authority

Pillar-cluster architecture & the citation flywheel

Rendering & crawlability

Can AI even see your JavaScript site?

AI search glossary

SEO, AEO, GEO, AIO & LLMO defined

GEO vs SEO vs AEO

What's the difference, in one table

Rank in Google AI Overviews

Get pulled into the AI Overview

Get cited by ChatGPT

Show up in ChatGPT & Perplexity answers

AI visibility tools

Track whether AI is citing you

E-E-A-T

Experience, expertise, authority, trust — who's behind the content

Surviving core updates

Diagnose a drop; recover the right way

URL cleanup & canonicals

301 vs 410, soft-404s, deindexing

Image & video SEO

Alt text, transcripts & YouTube as a citation source

Free GEO checker

Paste your HTML — instant AI-readiness audit

DIY vs a paid GEO service

When to do it yourself — and when paying is worth it

Content freshness

Keep facts, figures & claims current — a citation factor

How to cite ChatGPT & AI

APA, MLA & Chicago formats for AI sources

The agentic web: what to prepare for next

Where AI search is heading — browsers that act on pages, checkout inside the chat, crawlers you can license, and provenance labels in results. Six guides to get ahead of it.

Make your site agent-ready

Agentic browsers, WebMCP & the Lighthouse audit

How each AI engine picks sources

ChatGPT vs Perplexity vs AI Overviews vs AI Mode

Agentic commerce

ACP, AP2 & UCP: prepare your store for AI checkout

AI crawler control & licensing

Allow, block, or charge: pay-per-crawl & RSL

Content provenance

C2PA, SynthID & authenticity as a trust signal

Technical table stakes

The classic-SEO checklist AI visibility stands on

NW
OSINT researcher and founder; runs a network of ranked content & tools sites
Updated July 21, 2026

If you forget everything else

  1. SEO + GEO, not SEO vs GEO. Strong classic SEO is the foundation AI visibility is built on.
  2. AI search is a consensus engine that prefers earned media. Get the same true story onto multiple trusted, independent surfaces.
  3. Fact density plus information gain wins. Statistics, citations, expert quotes, and something original on every page. Rehashing the top three results earns nothing.
  4. Front-load the answer. First ~200 words, self-contained, carrying a number or a named entity.
  5. Challengers can win here. Lower-ranked pages gain the most, and AI citations have partly decoupled from the top-10 — you leapfrog by being more citable, not just bigger.

Questions people actually ask

How do I optimize my website for AI search?

Run five phases in order: audit what's blocking AI crawlers, front-load self-contained answers with real facts, generate the technical files (AI-crawler robots.txt, llms.txt, JSON-LD), build off-site consensus across the sources AI reads, and measure your share of voice. Each phase on this site has copy-paste Claude prompts.

What is AI search optimization called?

GEO (Generative Engine Optimization) and AEO (Answer Engine Optimization) are the common names, with AIO, LLMO, and "AI SEO" used interchangeably. See the glossary for how they relate — and note none of them replace classic SEO.

Is GEO just a rebrand of SEO?

Mostly it's an extension. The fundamentals — crawlable, fast, authoritative, well-structured pages — carry over, and Google says optimizing for its AI features is just SEO. GEO adds requirements around fact density, citation-friendly structure, and off-site consensus across the surfaces AI reads.

Do I need an llms.txt file to get cited?

No. An llms.txt is an agent-readiness convenience, not a citation lever — one large 2026 study found the vast majority of published llms.txt files were never requested. Chrome's Lighthouse now audits for it as agent-readiness, so ship it anyway — but spend your real effort on front-loaded answers and off-page consensus.

How long until this shows results?

Technical fixes (crawler access, schema, front-loaded answers) can influence citations within days to a few weeks because retrieval re-reads pages quickly. Off-page consensus compounds over months. Expect 40–60% citation churn week to week, so measure trends, not single checks.

Will Claude just make things up about my site?

It can if you let it. Keep it grounded: give it your real pages and a fixed entity fact sheet, ask it to work only from what you provide, and verify any statistic before you publish. Fabricated facts are exactly what the consensus engine is built to filter out.