An excerpt from LLM Mastery: How AI Recommends Brands for LLM Domination
This is the opening of the book, free to read with no email address asked for: the note that introduces its seven authors and the whole introduction, 791 words.
What does this excerpt contain?
The excerpt is the note that opens LLM Mastery: How AI Recommends Brands for LLM Domination and the whole introduction, printed exactly as the book has them.
It defines LLM domination as being named when AI models are asked about your market, and gives the authors' reason most markets are still uncontested: the industry is guessing.
- A Note Before We Begin
- 489 words
- Introduction: Domination Is Being the Answer
- 302 words
- In all
- 791 words, before Chapter 1
A Note Before We Begin
Seven of the market leaders in AEO and GEO wrote this book, and one of us is a digital avatar.
AI James Dooley is the digital avatar of James Dooley, the serial entrepreneur from Manchester who founded FatRank, PromoSEO and LLM Leads, and every word published under the AI James Dooley name is researched, written and refined by James Dooley personally.
Jabez Reuben is a leading pioneer in Generative Engine Optimisation and AI search visibility, founder of LLM Mastery and of The Blueprints, with over a decade spent reverse-engineering search algorithms. Disclosure done.
Mike Lovatt is a semantic SEO specialist and the director of M&B Marketing SARL, with around twenty years in search and a link building method built on corroboration: every placement is a witness statement the machines can cross-examine.
Paul Truscott is an award-winning qualified financial technical analyst turned search measurer. Paul Truscott coined Citation RSI, Entity Support and Resistance, Visibility Bollinger Bands and Visibility Drawdown, and has generated more than 150,000 leads for US home service businesses over six years.
Peter Jones has worked in search since 2008, generates more than 20,000 leads a year across a hyperlocal network of more than 4,000 towns, created the ECHO framework (Entity, Corroboration, Hooks, Output) and coined the terms Share of Answer and Entity Confidence.
Charles Floate is a British SEO consultant, educator and founder who started in search engine optimisation at the age of 12. Charles Floate founded and runs AI SEO Rainmakers, a fast growing AI SEO training community on Skool, founded SEO.Stream, the AI SEO publisher behind it, has trained more than 20,000 SEOs, and teaches every tactic from live tests.
Rollo Unden is a Swedish-British SEO entrepreneur, software developer and automation specialist. Rollo Unden is the founder of Apex Marketing, an automation-first SEO agency and software company founded in Guernsey, built Rapid Indexer, indexing software launched in 2025, and researches crawl discovery, indexing at scale and SERP reshaping.
Here is what unites the seven of us, and what this book runs on.
We test. Constantly. All seven of us share findings, compare results and interrogate every LLM on the market, every week, because opinions about AI search are worthless and experiments are not. The industry is drowning in conference-slide theory about how models pick brands. Almost none of it has been tested by the people repeating it.
Everything in this book comes from running the experiments: the same questions across every model, one variable at a time, logged verbatim, month after month. Where we state how AI recommends brands, it is because we watched it happen, repeatedly, across ChatGPT, Claude, Gemini, Perplexity and Grok.
The title is a promise with a condition. LLM domination, being the brand the machines name, is achievable in almost every market right now, but only for operators willing to test their way there. If you want beliefs, the industry has plenty. If you want the method, let's get into it.
The note introduces the seven authors, and each has a record on this site: AI James Dooley, Jabez Reuben, Mike Lovatt, Paul Truscott, Peter Victor Jones, Charles Floate and Rollo Unden.
Introduction: Domination Is Being the Answer
LLM domination means one thing: when the machines are asked about your market, they say your name.
Not your traffic. Not your rankings. Not your impressions. Those are the old scoreboards, and they still have their uses, but the game on top of them has changed. Customers now ask ChatGPT, Claude, Gemini, Perplexity and Google's AI surfaces what to buy, who to trust and which brand to choose, and the models answer with names. The named brands win conversations they never attended. The unnamed brands do not lose the argument. They are simply absent from it.
Domination in this era is share of the answers. Every relevant question, every model, every surface, your brand present, described accurately, recommended confidently. That is the entire objective, and it is measurable, buildable and, in most markets, shockingly uncontested.
Why uncontested? Because the industry is guessing.
Most GEO and AEO advice is theory repeated between people who have never tested it. Meanwhile the models themselves will show you exactly how they work, if you ask them systematically: what queries they fan out into, what sources they pull, why they chose the winners, what reasoning carried the verdict. The data is sitting in the answers, free, and almost nobody harvests it.
This book is the harvest method and the playbook built on it. How LLMs actually assemble answers. How the recommendation pipeline decides which brands get named. How the models differ and how consensus is built across them. The testing discipline, the five research datasets every campaign runs on, and then the build: training layer, retrieval layer, listicle frameworks, entity foundations, off-page signals, the five money queries, and the measurement system that tracks domination month by month.
The machines are answering questions about your market today, with or without you in the answer.
Let's make it with.
The introduction is summarised on its chapter page, and LLM domination has its own glossary entry.
What comes after the introduction?
Chapter 1: How LLMs Actually Answer comes next.
Every LLM answer is assembled from two layers, memory and retrieval, joined by a synthesis step, and a brand's visibility depends on all three. Memory is what a model absorbed in training and updates only when the model retrains; retrieval is the live search it runs mid-answer, and a change to what it can fetch shows up in answers within days.
The chapters page lists the Introduction and 14 chapters, each with its first sentence. The full text is in the ebook, and Where to buy the ebook lists each store.