Link Building in the Era of AI
Get the book

Link building in AI search: a glossary

Corroboration

What is Corroboration?

Corroboration is the agreement between independent sources on the same claim about an entity, and it is what raises a retrieval system's confidence that the claim is true.

Who uses the term Corroboration?

Corroboration is used across semantic SEO and entity optimisation, and is central to the ECHO framework created by one of this book's authors.

The term is borrowed from evidence law and from information retrieval, where it carries the same meaning: one source asserts, several agreeing sources establish.

How does Corroboration differ from related terms?

Corroboration is agreement across sources. Consistency is agreement within one source. A site can be perfectly consistent with itself and still uncorroborated by anything else, which is the condition most brands are in.

Authority is a score attached to a domain. Corroboration is a property of a claim, and it is the property AI systems are actually testing.

What does Link Building in the Era of AI Search Engines say about Corroboration?

The book's strictest line on corroboration is about wording, not volume: the machine counts agreement, and agreement requires the same claim in the same words appearing everywhere. It devotes chapters 4, 5 and 6 to corroboration across social platforms, directories, and press and guest placements, on the grounds that the machine is checking whether the witnesses agree.

Read about the book's argument.

Related terms

  • Witness statement

    A witness statement is this book's term for a link once the surrounding page is read rather than counted: the value sits in what the page says about you, not in the link itself.

  • Corroboration ledger

    A corroboration ledger is a running record of which independent sources assert which claims about an entity, kept so that link work can be measured as evidence gathered rather than as links acquired.

  • Entity development

    Entity development is the work of making a brand resolvable as one clearly defined thing, so that mentions of it accumulate onto a single record rather than scattering across several ambiguous ones.

  • Machine-read testimony

    Machine-read testimony is the body of statements about a brand that AI search systems actually read and weigh, as distinct from the links they were once assumed to count.

All 14 terms Link Building in the Era of AI Search Engines