Grounded guides on generative engine optimization: what's real, what's cited research, and what's honestly still unknown. No invented stats, no guaranteed formulas.
Training crawlers are blocked twice as often as AI search crawlers. Who blocks what, from an October 2026 crawl of the top 10,000.
16.1% of top websites serve an llms.txt, and about half of those files break the format.
JSON-LD, Organization markup and visible text on 486 homepages, read without JavaScript.
The lines crawlers silently skip, Crawl-delay and Noindex leftovers, and early Content-Signal adoption.
Robots.txt, llms.txt and structured data on 18 vendors’ homepages, ours included.
Claude’s three crawlers, the three search providers Anthropic lists, and a plan.
How Perplexity finds and footnotes pages, and how to become one of them.
Grounding in Google Search, and the Google-Extended setting most sites misread.
Google’s own requirements, query fan-out, and the controls that keep you out.
Copilot asks Bing a few words and cites what comes back. What that means for you.
A practical experiment design for GEO, with the sample sizes you need to detect a real lift.
Monthly Google searches for nine AI assistants in the US and UK, and what it means for where to be visible.
US and global search demand for GEO, AEO, LLM SEO and AI-visibility terms, with the full table.
Search demand for 16 AI crawler names: what site owners look up, and the search crawlers they overlook.
Difficulty, volume, CPC and traffic potential for 29 GEO and AEO terms, easiest first.
Confidence intervals and minimum detectable changes for AI visibility scores, from 5 answers to 400.
Exact formulas for AI visibility rate, share of voice and average position, with a worked example.
A copy-ready monthly AI visibility report for clients: sections, metrics, tables and honest wording.
Organic traffic, domain rating and branded search demand for 18 AI visibility vendors, ourselves included.
300 real Google questions about AI search visibility, with volumes, grouped into eight themes.
What GEO is, how it differs from SEO, why it exists, and what's actually in your control today.
What a rank tracker measures vs. what an AI citation checker measures, and where the two genuinely overlap.
Honest, practical tactics — and a clear-eyed look at what nobody can actually guarantee.
What a realistic AI buyer question looks like, and why testing one prompt isn't enough.
How location-qualified AI answers work, grounded in real local-SEO research.
What schema.org markup actually does for AI crawlers, versus unproven SEO folklore.
What the real usage and zero-click data says about how buyers research now — and what it doesn't.
What's publicly documented — and what's genuinely not knowable from outside the model provider.
A practical, prioritized checklist — starting with knowing your baseline, not chasing tactics.
Manual spot-checks, one-off checks and weekly tracking compared, and what to look for in a tracking tool.
GPTBot, OAI-SearchBot, ClaudeBot and Google-Extended do different jobs. What blocking each one changes, and a robots.txt to copy.
The format, who reads it today, what to put in yours, and why it isn’t the reason an assistant recommends you.
A custom channel for AI assistants in ten minutes, the regex to paste, and the AI traffic GA4 will never show you.
GEO, AEO, share of voice, grounding, AI Overviews, GPTBot, llms.txt and more, each defined in a sentence or two.
Whether your robots.txt lets GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot and 14 more in.
Check a site’s llms.txt against the format, or build one from your key pages.
JSON-LD, noindex and nosnippet, and server-rendered text, as a crawler sees them.
The questions your buyers ask AI assistants, grouped by intent.
Seven tools on entry price, free option, coverage and what you track, sourced and dated.
An enterprise platform and a free-to-start one, side by side, including where Profound goes further.
Price, models, prompts and cadence, from both pricing pages.
The cheapest daily tracker and a free plan, compared honestly.
Credits across 11 platforms against a free plan, with the arithmetic.
Per-seat pricing, prompt limits and coverage, side by side.
A 514-million-prompt index and metered checks against a flat plan.