Learn the AI web
A plain guide to answerability, AEO and GEO: the concepts, the metrics, and how to use them. Written to stand on its own and be useful even if you never run the tool.
Tutorials
Find why an AI cannot read your page
A few minutes to see what a non rendering AI crawler actually gets, and why.
Fix a page that ranks but is not answered
When a page gets search traffic but never gets cited by AI.
Trim crawl waste on your site
Stop bots spending their budget on redirects, errors and duplicates.
How to read a Topkay audit
What the score, the lenses and the fixes mean, and where to start.
Concepts
What is ARMX
Agent Readiness and Machine eXperience: the five things every machine does with your site, and the one score that measures them.
The origin and its many viewpoints
One source, read differently by browsers, crawlers, AI agents, and a growing list of tools, plugins and apps.
What crawl budget is
The finite attention a crawler gives your site, and how it gets wasted.
What a digital twin of a website is
One graph of your site, inspectable from every reader's view.
How AI search differs from classic search
Classic search gives you links to click. AI search gives you an answer, assembled from sources.
How retrieval works: chunks and embeddings
How an AI finds the right passage on your page, and why the passage is the unit.
Trust, and why AI answers care about it
Engines repeat sources they trust. What builds that trust on a page.
Server rendering and client rendering
Where your page is built decides who can read it.
Share of voice in AI answers, and the honest take
How often AI names your brand. Worth tracking, but downstream of the page-level work.
Zero-click, when the answer needs no visit
The answer appears in the result or the chat, and no one clicks through.
The Topkay crawl format, an open standard for crawls
A portable format for a website crawl: pages, links and scores, built on JSON Graph Format.
Metrics
What internal PageRank is
How link equity flows through your own site, and why buried pages lose out.
What fan-out coverage is
The questions an AI expands your topic into, and which your page leaves unanswered.
Core Web Vitals, plainly
The three numbers that measure how a page feels to a person.
Definitions
What is answerability
How easily an AI answer engine can find and lift the answer from your page.
SEO, AEO and GEO, the difference
Ranking pages, being the answer, being cited by models. Three jobs, not three names.
What cloaking is
Showing a crawler something different from a person, and why it backfires.
What a redirect chain is
One URL that bounces through several before it resolves, and the cost.
What structured data is
Machine readable facts about a page, added as JSON-LD, so engines can use them directly.
What a canonical URL is
The one address you want engines to treat as the real version of a page.
What robots.txt does
A file at your root that tells crawlers which paths they may fetch.
Why Topkay does not use llms.txt
A proposed index for models that we deliberately do not check, recommend or generate.
HTTP status codes, the ones that matter
The numbers a server returns, and what each tells a crawler.
What an orphan page is
A page nothing else links to, so it is hard to reach and inherits no equity.
What a sitemap is
A list of your URLs that tells engines what exists, and how to keep it honest.
What anchor text is
The visible words of a link, which tell engines what the target is about.
What hreflang is
How you tell engines which language and region a page is meant for.
How-tos
Why AI cannot see content behind JavaScript
Non-rendering crawlers get the shell, not the content, and how to fix it.
How to make a page more answerable
Five changes that help an engine lift your answer.
Titles and meta descriptions
The two pieces of text that decide how your page looks in a result.
Headings and document structure
How to structure a page so people and engines can follow it, and lift from it.
Reading server logs to see what crawlers do
Your access log is the ground truth of how bots actually crawl you.