Cover of Winning AI Search by LIPAI WANG

The book

Winning AI Search

The Indie Developer's Playbook for Getting Cited by Google AI Overviews, ChatGPT, Perplexity, and Claude in 2026

by LIPAI WANG · 16 chapters · 18 diagrams · ~20 worked formulas

Get it with the course →Read the free articles

Coming to Kindle shortly. Course members get the EPUB and PDF at no extra cost when it lands.

Search stopped being one thing

A single question now fans out across at least five independent surfaces, each with its own crawler, its own signals, and its own idea of who deserves credit. Optimizing for one buys you almost nothing on the others. Three numbers explain why the old playbook solves for the smallest slice of the pie:

58%
of searches end in zero clicks
11%
overlap between ChatGPT and Perplexity citations
47%
of AI Overview citations rank below position 5

Together they mean ranking and being cited have split into two economies. The book is a field manual for the second one — the one a solo developer can still win cheaply, because the correlation between domain authority and citations collapsed from r=0.55 to r=0.18. A clean new domain can get cited the week it publishes.

Inside the book

I

The New Search Reality

Five surfaces, not one. Why ranking and citation split into separate economies — and a scoring rule for picking the two surfaces worth your hours.

II

The Schema Trust Layer

Why marked-up pages earn roughly 73% higher citation rates, the copy-paste JSON-LD to earn it, and the validation workflow that stops schema failing silently.

III

Positioning and Penalties

The protection-racket trap, the five triggers behind site-wide Scaled Content Abuse demotion, and why directory-template clones get suppressed.

IV

The Technical Field Guide

The Cloudflare default that silently blocks GPTBot and ClaudeBot, the SPA fallback that serves your homepage at /robots.txt, and the five-minute IndexNow win.

V

Measure What Matters

Read a Search Console panel like a diagnostician, and run a measurement loop that attributes every change you make.

Diagrams, not decoration

Every chapter carries figures and formulas you can run against your own site. Two of the eighteen:

One user question fanning out to five independent search surfaces — Google blue links, AI Overviews, ChatGPT Search, Perplexity, and Claude/Gemini — each with its own crawler and citation pattern.
Chapter 1 — the five surfaces, and why coverage doesn't transfer between them.
Five trigger mechanisms feeding a logistic risk gate: cross roughly two or three triggers and the output flips to site-wide demotion across the whole domain.
Chapter 9 — why two or three triggers is enough to demote an entire domain.

Book or course?

The book is the written foundation — read it end to end in an evening. The course is the living version: the same material as lessons, plus new modules, the companion tools, and every future update. Course members get the book included.

See what's in the courseBrowse the curriculum