Glossary of LLM Mastery: How AI Recommends Brands for LLM Domination
The book defines 27 terms of its own and credits 7 more to two of its authors; each entry below gives its definition, or its owner, and the page that holds it.
LLM domination means one thing: when the machines are asked about your market, they say your name.
How many terms does the glossary hold?
The glossary holds 34 entries: 27 terms the book defines and 7 terms it credits to two of its authors.
12 have a page of their own, 15 live in the chapter that defines them, and the 7 credited terms link to their authors' own definitions. The book's own terms come first, in the order its chapters define them, then the credited terms.
- LLM domination
- LLM domination is the book's name for being the brand AI models name when they are asked about your market. Defined in Introduction: Domination Is Being the Answer
- Training layer
- The training layer is what an AI model absorbed about your brand in training, which it recalls without searching and which changes only when the model is retrained. Also called memory layer. Defined in Chapter 1: How LLMs Actually Answer
- Retrieval layer
- The retrieval layer is the live search an AI model runs while it answers, pulling current pages, reviews, rankings and coverage into its response. Defined in Chapter 1: How LLMs Actually Answer
- Synthesis step
- The synthesis step is where an AI model writes its answer, weighing what it remembers and what it retrieved against the question and choosing which brands to include. Defined in Chapter 1: How LLMs Actually Answer
- Recommendation pipeline
- The recommendation pipeline is the book's four-stage account of how an AI answer comes to name a brand: recognise, evidence, confidence, name. Owned by Chapter 2: How AI Recommends Brands, which defines it:
A brand recommendation is the output of a pipeline, and the pipeline has four stages: recognise, evidence, confidence, name.
- AI consensus
- AI consensus is the state where every AI model, on every surface, reaches the same conclusion about your brand. Owned by Chapter 4: AI Consensus: Make Every Model Agree on You, which defines it:
AI consensus is the state where every model, on every surface, reaches the same conclusion about your brand, and it is the real asset this entire discipline builds.
- Consensus rate
- Consensus rate is how often AI models agree on including your brand and describe it the same way, measured across the questions that connect to revenue and the models that matter. Also called agreement rate. Defined in Chapter 4: AI Consensus: Make Every Model Agree on You
- Test Test Test method
- The Test Test Test method is the book's testing discipline: a fixed bank of buyer questions run identically on every model, logged verbatim and repeated at least monthly. Owned by Chapter 5: The Test Test Test Method, which defines it:
Every author of this book runs the same discipline, shares the findings and holds the same rule: no claim survives unless the models confirm it.
- Prompt bank
- The prompt bank is the fixed set of questions real buyers ask about a market, worded once and run unchanged on every model at every test. Defined in Chapter 5: The Test Test Test Method
- The five datasets
- The five datasets are the research the book says every logged AI answer contains: fan-out queries, reasoning data, cited sources, winners and sub-categories. Also called five research datasets. Owned by Chapter 6: Campaign Research: The Five Datasets That Decide GEO and AEO Strategy, which defines it:
Every LLM answer you log contains five datasets, and together they are the complete research layer of a GEO and AEO campaign.
- Fan-out queries
- Fan-out queries are the sub-queries an AI model generates and researches from one user prompt before it writes a single answer. Defined in Chapter 6: Campaign Research: The Five Datasets That Decide GEO and AEO Strategy
- Reasoning data
- Reasoning data is the visible reasoning an AI model shows after gathering its sources, which records the criteria it applied when it preferred one brand to another. Defined in Chapter 6: Campaign Research: The Five Datasets That Decide GEO and AEO Strategy
- Cited sources
- Cited sources are the pages AI answers cite for your market, which the book treats as the list of places to get listed and described accurately. Also called citation dataset and citation list. Defined in Chapter 6: Campaign Research: The Five Datasets That Decide GEO and AEO Strategy
- Sub-categories
- Sub-categories are the narrower territories that AI answers reveal inside a market, such as budget options, premium picks and specialist niches, and the book treats each one as winnable. Defined in Chapter 6: Campaign Research: The Five Datasets That Decide GEO and AEO Strategy
- Liftable
- Liftable describes content an AI model quotes cleanly: the question as the heading, the direct answer first, one claim per sentence and specifics a machine can cite. Also called machine-liftable. Defined in Chapter 8: Winning the Retrieval Layer
- Listicle frameworks
- Listicle frameworks are the book's rules for ranked lists that AI models lift intact: numbered positions, stated selection criteria, checkable facts per entry, exact brand names and honest breadth. Owned by Chapter 9: Listicle Frameworks for AI Crawlers, which defines it:
A listicle works for AI crawlers when its structure survives extraction.
- Entity foundations
- Entity foundations are the four permanent practices that let AI models attach evidence to the right brand: be resolvable, publish the entity home, sweep the contradictions and guard the name. Owned by Chapter 10: Entity Foundations for LLM Visibility, which defines it:
Every tactic in this book leaks without entity foundations, because evidence only counts when the machines can confidently attach it to you.
- Entity home
- An entity home is one page on your own site that states your complete record in plain words a machine can quote, with the labels and profile links that tie your accounts into one identity. Defined in Chapter 10: Entity Foundations for LLM Visibility
- Off-page signals
- Off-page signals are what other sites say about a brand, and the book measures them as the share of a market's trusted surfaces that state the brand's facts accurately and in agreement. Owned by Chapter 11: Off-Page Signals LLMs Trust, which defines it:
Measured properly, off-page in the AI era is not a link count. It is the share of your market's trusted surfaces that state your facts, accurately, in agreement.
- Brand sheet
- The brand sheet is the set of facts every off-page placement carries: the exact name, the agreed description and current facts. Owned by Chapter 11: Off-Page Signals LLMs Trust, which defines it:
Every placement carries the brand sheet: exact name, canonical description, current facts, so witnesses tell the same story.
- The five money queries
- The five money queries are the question shapes that brand traction creates: competitor, comparison, alternative, pricing and review queries. Also called money queries and five query shapes. Owned by Chapter 12: Owning the Five Money Queries, which defines it:
Brand traction spawns five query shapes, and dominating a market means covering all five, because buyers are won at discovery and lost at the decision.
- Presence rate
- Presence rate is how often AI models name your brand at all, across your money questions and models. Owned by Chapter 13: Measuring LLM Domination, which defines it:
Presence rate. Across your money questions and models: how often are you named at all.
- Position and framing
- Position and framing record where and how a brand is named: first mention or afterthought, flat recommendation or hedged inclusion, your own agreed wording or an approximation. Owned by Chapter 13: Measuring LLM Domination, which defines it:
Position and framing. When named, where and how: first mention or afterthought, flat recommendation or hedged inclusion, described in your canonical language or approximated from scraps.
- Fan-out coverage
- Fan-out coverage is the share of a market's harvested fan-out terms on which your brand appears when each is probed directly. Owned by Chapter 13: Measuring LLM Domination, which defines it:
Fan-out coverage. Of the fan-out terms harvested for your market, what share show you present when probed directly.
- Citation share
- Citation share is how many of the sources AI answers cite carry your brand, and carry it accurately. Owned by Chapter 13: Measuring LLM Domination, which defines it:
Citation share. Of the sources the answers cite, how many carry you, accurately. This is the supply-side metric: it moves before presence moves, which makes it the early indicator that the campaign is landing.
- Domination playbook
- The domination playbook is the book's seven-step sequence for LLM domination, from baselining the answers to running a monthly interrogation. Owned by Chapter 14: The Domination Playbook, which defines it:
LLM domination is built in a sequence, and the sequence exists because each stage stops the next one leaking.
- Monthly interrogation
- The monthly interrogation is the last step of the book's playbook: re-running the tests every month, re-harvesting the five datasets, feeding the gaps and steering by the trendline. Owned by Chapter 14: The Domination Playbook, which defines it:
Re-test, re-harvest, read the reasoning, feed the gaps, track presence, framing, consensus, fan-out coverage and citation share, and steer by the trendline.
- Citation RSI
- Credited to Paul Truscott. The book names it without defining it; the definition is on paultruscott.com.
- Entity Support and Resistance
- Credited to Paul Truscott. The book names it without defining it; the definition is on paultruscott.com.
- Visibility Bollinger Bands
- Credited to Paul Truscott. The book names it without defining it; the definition is on paultruscott.com.
- Visibility Drawdown
- Credited to Paul Truscott. The book names it without defining it; the definition is on paultruscott.com.
- The ECHO Playbook
- Credited to Peter Victor Jones. The book gives only its name and expansion,
the ECHO framework (Entity, Corroboration, Hooks, Output)
; the definition is on echoplaybook.com. Also called ECHO framework. - Share of Answer
- Credited to Peter Victor Jones. The book names it without defining it; the definition is on echoplaybook.com.
- Entity Confidence
- Credited to Peter Victor Jones. The book names it without defining it; the definition is on echoplaybook.com.
Where do the definitions come from?
Every definition comes from the manuscript: each entry names the chapter that defines it, and the entry's own page or chapter quotes the book's sentence word for word.
The terms LLM Mastery: How AI Recommends Brands for LLM Domination defines, quoted from the book, and the terms it credits to its authors, linked to their own definitions.
A term has its own page when the book defines it in a sentence that is not the opening of a chapter named after it, and another part of the book uses it; otherwise the chapter that defines it owns it.
Used in the book but not defined
The book uses six more terms without defining them, so this glossary gives them no definition.
- Grounding chunks
- GEO, defined on aeogeollmseeding.com
- AEO, defined on aeogeollmseeding.com
- Head term
- Closed-book questions
- Money questions
Which terms belong to the authors rather than the book?
The book credits Paul Truscott with Citation RSI, Entity Support and Resistance, Visibility Bollinger Bands and Visibility Drawdown, and Peter Victor Jones with The ECHO Playbook, Share of Answer and Entity Confidence; it names these terms without defining them, and their definitions live on the authors' own sites.
- Paul Truscott: definitions on paultruscott.com
- Peter Victor Jones: definitions on echoplaybook.com