The Shift
Why being found and being recommended stopped being the same job.
- 1
Search Didn't Die. It Moved.
What actually changed, and why a page that ranks well can still be absent from every answer.
- 2
Selection: How AI Actually Builds an Answer
The mechanism. An engine retrieves passages, weighs sources and composes one answer naming a few brands. Everything else in the book follows from this.
- 3
The Author Paradox: What 1,000,000 Websites Reveal
The evidence chapter. 94% of firms displayed their qualifications and 8% made them machine-readable, which is the legibility gap in one statistic.
- 4
The Two Games: AI's Memory vs AI's Search
Memory answers come from training and move on training timescales. Retrieval answers come from a live index and move in weeks. They need different work.
Diagnose
Find out where you actually stand before you change anything.
- 5
Run Your Visibility Baseline
The locked prompt set, run properly, so you have a number to compare against later.
- 6
The Competitor Replacement Test
Your baseline says whether AI picks you. This says why it picked someone else, which is the part that points at the fix.
- 7
Read Your Baseline: Not Seen, Sometimes Seen, or Winning
Three states, three different first moves. Find yours and skip the work that will not help you yet.
Foundation
The mechanical layer. Mostly one-off, mostly undone by your competitors.
- 8
Crawler Access: The Right Door for Each Engine
Search bots govern citations and training bots govern memory. They are different user agents and blocking the wrong one is a common, silent cause of invisibility.
- 9
llms.txt: Cheap Insurance, Not a Magic Bullet
What the file does, what it does not do, and the honest reason to spend five minutes on it anyway.
- 10
Schema and Structured Data, Without the Myths
What the evidence actually supports, what it does not, and the blocks worth adding as disambiguation hygiene.
- 11
Entity Clarity and the Brand Entity Block
An engine can read every word on your site and still not know what you are. This is how you tell it.
- 12
The 15-Minute Wikidata Walkthrough
A free entry in the knowledge base most engines draw on, done properly, in a quarter of an hour.
Extraction
Writing so a machine can lift your answer cleanly and attribute it to you.
- 13
Answer-First: Write the Answer, Then the Page
Traditional copy builds to a reveal. Retrieval rewards the opposite, and measured citations cluster near the top of a page.
- 14
The Five Blocks
Five AI-ready content blocks to copy, fill in and place directly into your pages.
- 15
The Perfect Page, and the Page That Fails
The blocks assembled into one complete page, next to its opposite, so the difference is visible.
- 16
Query Fan-Out: Why One Page Is Never Enough
An engine splits one question into many hidden sub-queries. Covering a decision beats perfecting a page.
- 17
One System, Any Industry
The same structure applied three times with different inputs, so it scales past the worked example.
- 18
E-commerce: Selling Where AI Shops
Product data is a second surface with its own rules, and it is the one retailers most often leave unread.
Reinforcement
Third parties decide what AI believes about you.
- 19
Why Third Parties Decide, and How to Find Yours
Foundation and Extraction get you considered. This is the layer that gets you chosen, and it is decided mostly off your own site.
- 20
The Mention Engine: Profiles, Reviews, and Press
The actual machine: what to set up, what to send, and what finished looks like for a small business.
- 21
Become the Stat Source
Own a number. Publish original data about your niche and mentions stop being something you chase.
- 22
E-E-A-T and Author Authority
Credentials a human can see and a machine cannot are worth nothing here. How to make expertise legible.
- 23
Topical Authority and Content Hubs
Whether you cover your subject or have two shallow pages, and why that decides fanned-out queries.
- 24
Off-Site Signals: Where AI Actually Gets Its Sources
The wider web of places your brand lives, and which of them engines lean on most.
Track and repair
Measure it as a rate, separate noise from drift, and fix what slips.
- 25
Case Study: How FileSeal Out-Recommended DocuSign in 90 Days
One company, the full sequence, with the scores and the dates. The worked example for everything in Parts 3 to 5.
- 26
Volatility and Drift: Why One Audit Is Never Enough
Same-day runs of the same prompt overlap by roughly a third, so a single run means nothing and rates are the unit.
- 27
Build Your Tracking System
The minimum viable system: same questions, same way, on a schedule, recording rates.
- 28
Reading Your Results, Engine by Engine
The six engines behave differently. What each one rewards and how to read a result that only moves on some of them.
- 29
Objective Measurement: Server Logs and AI Referrals
Two measurements your own systems can produce, and the honest limit of what each proves.
- 30
When AI Gets You Wrong: The Repair Playbook
Wrong prices, features you do not have, a service you stopped offering. What to change, and in what order, to correct it.
- 31
Insight to Fix to Result
Closing the loop, and a straight word on where a tool helps and where it does not.
Execute
Plans you can actually run, at two different sizes.
- 32
The 7-Day Starter Sprint
One focused task a day, from nothing to a baseline, one optimised page and a re-test booked.
- 33
The 30-Day AI Visibility Plan
One week per layer, in order, with a deliverable at the end of each.
- 34
The Priority Matrix, the 8 Mistakes, and the Troubleshooting Tree
How to decide what to fix first, the mistakes that quietly sink most efforts, and what to do when nothing moves.
- 35
The Agentic Shift: Optimising for AI That Acts, Not Just Answers
Agents shortlist and act, and the user never sees what was filtered out. What changes when that is the buyer.
- 36
Choose Your Path
Do it yourself, do it with a tool, or have it done. The honest trade-offs of each.
The toolkit
Everything in one place, plus your access code.
The baseline and tracking sheet, the thirty-prompt query bank, the pre-publish checklist, the self-scorecard, the priority matrix, the five core templates, the file pack, four worksheets, the Notion templates and the glossary. The access code that unlocks the digital versions is printed here.
Start with Chapter 2
It explains how an AI assembles an answer and decides who to name. Everything after it depends on that mechanism, so it is the fairest test of whether the book is for you.