Chapter 10: Coverage the Machines Cannot Use
A large share of the coverage brands pay for and celebrate contributes nothing to AI visibility, and knowing which placements are inert saves more budget than any new tactic.
So Chapter 10 opens.
THE BOOK
What does this chapter argue?
The chapter lists eight kinds: the bare mention, the inconsistent mention, the paywalled paragraph, the mention inside an image, the syndication echo, the unattributed quote, the sponsored label without substance and the dead page. Most brands find few usable placements.
All of them fail one test: could a machine extract a complete, attributable, checkable statement about the brand, in its canonical form, without help? Auditing three years of coverage against that test usually shows that a minority of placements carry usable facts, and that minority is rarely the biggest audiences.
SOURCE
Summarised from the text of Chapter 10. Every sentence on this page traces to the chapter's own lines; the facts behind the site are on the record.
How does the chapter open and close?
It opens with the line quoted at the top of this page, and runs to two printed subheadings before it closes.
CLOSING LINE
If a machine cannot lift a fact from it, it is a clipping, not evidence. Report it that way.
What are the chapter's sections?
The chapter has two printed subheadings, in this order:
- What the inert categories have in common
- Auditing what you already have
Which terms and people does the chapter name?
The terms it uses from the glossary:
- The liftable test
PEOPLE NAMED
The chapter names no person, company or AI system.
Where does it sit in the book?
ALL CHAPTERS