How to Get Found in AI Answers

An entire vocabulary appeared almost overnight — AEO, GEO, llms.txt, MCP, agent cards — promising businesses a way into AI-generated answers. Most of it arrived faster than any major platform officially endorsed it.

This book checks that vocabulary against the only sources that actually matter: what Google, OpenAI, and Microsoft themselves publish about how their own systems work, and, where available, real independent measurement of whether businesses following popular advice are seeing any result at all.

Chapter 2 covers what Google's own developer documentation says about appearing in AI Overviews and AI Mode — one statement that undercuts a large portion of the AEO industry's central pitch. Chapter 3 separates OAI-SearchBot, GPTBot, and ChatGPT-User into their actual distinct roles, and explains what blocking or allowing each one does and doesn't do. Chapter 4 examines the llms.txt recommendation against a large-scale study of 137,210 domains: 97 percent of all llms.txt files received zero requests during the entire study month.

Chapter 5 covers what sitemaps, Search Console, and IndexNow actually do — and what they don't guarantee. Chapter 7 covers the one real, available measurement: how to track ChatGPT referral traffic using utm_source=chatgpt.com, turning a mostly speculative subject into something testable with real evidence.

Nine chapters. Every claim verified directly against Google's, OpenAI's, and Bing's own primary documentation. Freshness date: August 8, 2026. By Pop Buchanan.