The GEO knowledge base

GEO or AEO? The two acronyms name the same work — optimizing content to be cited in AI answers. We define both, trace where each came from, explain why we treat GEO as the umbrella term, and show why the distinction some draw between them does not hold up in practice.

LLMs build a brand's picture from corpus consensus across many sources, not from your own self-description. How to organize your brand entity — canonical name, Organization schema, sameAs, Wikidata — and audit the consistency models actually see.

E-E-A-T in an AI context without the marketing myths: how to build author pages, bios with sameAs, and Person schema, what Google actually says about people-first authorship, and why real expert authors beat anonymous content. Every number carries a source and a date.

The Bing index is the official condition for visibility in Microsoft Copilot and still an important grounding channel for ChatGPT. Here is how to add your site to Bing Webmaster Tools step by step and confirm it is indexed.

A precise definition of AI share of voice and how it is measured: the two calculation methods, why mentions and ghost citations are separate KPIs, per-engine and per-market measurement, and what actually moves the number — built on 2025–2026 studies from Semrush, Ahrefs and Muck Rack.

The keyword became a prompt: longer, contextual, multi-intent. How to collect prompts from PAA, forums, sales calls, and tools, how to map them to content by intent, and why keyword research no longer covers how customers ask.

Earned media accounts for ~84% of citations in AI answers, and brand mentions correlate with visibility more strongly than backlinks. What the Muck Rack, Ahrefs and Omniscient data shows — and what it means for digital PR for AI.

61.7% of AI citations never name the brand. What a ghost citation is, how brand mentions differ from citations, why ChatGPT and Gemini fail in opposite ways, and what it means for measuring AI visibility.

How ChatGPT search selects the sources it cites: the role of the Bing and OAI-SearchBot indexes, freshness, semantic match, content structure, and brand mentions — with data from Ahrefs, Semrush, and Muck Rack. Plus a list of actions that realistically make sense for a brand.

What does LLM mean? A plain-English guide for marketers: how large language models work, training vs. grounding, tokens and context windows, why models hallucinate — and what it all means for your brand's content.

Perplexity handled 780 million queries in May 2025 and cites sources in every answer. Learn how the answer engine works, how it differs from ChatGPT and Google, and how to check whether it cites your website.

A step-by-step method for monitoring brand visibility in AI answers: building a prompt set, separating brand mentions from ghost citations, measuring per engine, choosing a cadence, and deciding between DIY and AI visibility tools — built on 2025–2026 studies from SparkToro, Semrush, Ahrefs and seoClarity.

A writing workshop for AI search: BLUF from the US Air Force handbook, the lift-out test for quotable sentences, before/after rewrites, intent-format match for lists and tables, and a 7-point publishing checklist — built on 2024–2026 citation studies.

Every AI provider sends multiple crawlers with different jobs: model training, AI search indexing, and on-demand page fetching. We break down what each robots.txt token does, what blocking it costs, and which rules to use for three strategies.

A step-by-step llms.txt implementation guide with syntax examples — plus the honest part: 97% of published files get zero requests, and no major LLM provider officially reads them.

AI SEO, GEO, AEO, LLMO, AIO — five names, one discipline. We define AI SEO, explain why GEO is the canonical term, show how it differs from classic SEO, and break down the five work streams the practice covers, from AI crawler access to measuring AI share of voice.

A living benchmark of AI search statistics: engine adoption, AI Overviews frequency, CTR impact, and ChatGPT referral traffic — every number with a source and measurement date, reviewed quarterly.

Every major AI citations study in one place: the GEO paper (KDD 2024), the C-SEO Bench replication (NeurIPS 2025), and the large observational datasets from Ahrefs, Semrush, Profound, Muck Rack, Wix and Kevin Indig. Each finding sorted into proven, correlated, or a bet — with a study-by-study evidence table.

Google says no special schema is needed for AI Overviews, FAQ rich results are fully deprecated as of May 2026, and the only large causal test found no citation uplift from adding JSON-LD. Here is what structured data actually does for AI search — and what agencies should stop promising.

The full seven-step process to optimize a website for AI search — from a robots.txt and WAF audit for GPTBot, OAI-SearchBot, PerplexityBot, and ClaudeBot, through JavaScript rendering and answer-first structure, to Bing Webmaster Tools and citation monitoring. Built on official documentation and studies, not promises.

The GEO paper's 'up to 40%' is the most quoted number in generative engine optimization research — and the least understood. We walk through Aggarwal et al. (KDD 2024) tactic by tactic, with exact figures, then through C-SEO Bench (NeurIPS 2025), the independent replication that confirmed almost none of it.

Training vs grounding, the OAI-SearchBot and Bing indexes, ghost citations and unstable citation patterns. A data-backed breakdown of how ChatGPT, Gemini and Copilot pick the sources they cite, built on 2024–2026 studies and platform documentation.

12 checks that decide whether generative engines can cite your website — from robots.txt and JavaScript rendering to dates, authors, and structured data. Each check: what to test today and how to fix it.

AI Overviews rewrite the rules of Google visibility: a top-10 ranking no longer guarantees clicks. We break down how Google generates its AI summaries, which searches trigger them, and what Pew, Ahrefs, and Semrush data say about the traffic impact.

The complete guide to GEO: the definition from the KDD 2024 paper, how generative engines pick their sources, what changes versus SEO, a six-step implementation process, measurement with three separate KPIs, and five myths dismantled with primary data.

Two disciplines, one goal — and completely different rules. Where SEO ends and GEO begins.