Source Attestation
The process by which an AI model verifies the authority and truthfulness of a source before citing it in an answer.
What is Source Attestation?
Source attestation describes the set of checks an answer engine applies before it treats a page as trustworthy enough to cite. Engines do not publish these checks in detail, but their behaviour and public guidance point to consistent themes: whether the page is on a domain with an established reputation for the topic, whether the claims are specific and consistent with other retrieved sources, whether the content has a clear author or organisation behind it, whether it is current, and whether it presents information plainly rather than as promotion.
In practice attestation is a combination of retrieval ranking and model judgement. The retrieval layer scores pages by relevance and by domain-level signals; the model then reads the passages and tends to prefer those that agree with each other and state facts directly. A page that contradicts the consensus of other sources, or that makes claims without support, is less likely to be quoted even when it is retrieved.
For brands the actionable version is to make every fact easy to verify: name the organisation and author, date the content, cite primary sources, keep numbers consistent across the site and third-party profiles, and use structured data (Organization, Product, Article, FAQPage) so the entities and claims are machine-readable.
Why it matters for AI search
Being retrieved is necessary but not sufficient; the engine still has to decide the page is worth citing. Attestation signals are what tip that decision. A brand whose facts are consistent everywhere, whose pages are clearly authored and dated, and whose claims match what other trusted sources say gets cited more and hallucinated about less. Inconsistent or vague information does the opposite, and it is one of the most common reasons an otherwise strong page is passed over.
Related terms
Citation Intelligence
The analysis of which specific sources and pages an AI model chooses to reference when answering a query.
Hallucination
When an AI model generates factually incorrect information confidently. Monitoring hallucinations is critical for brand safety.
Knowledge Graph
A network of entities and their relationships that AI models use to provide structured, factual answers.
RAG (Retrieval-Augmented Generation)
A technique that allows LLMs to pull in real-time information from external sources (like a website) before generating an answer.
E-E-A-T
Experience, Expertise, Authoritativeness, Trustworthiness. Google's quality criteria that AI models also consider when evaluating content authority.
Frequently asked questions
What signals help an AI engine trust a page?+
Clear authorship and organisation identity, dated and current content, specific claims backed by primary sources, consistency with other retrieved pages, relevant structured data, and a domain with a track record on the topic.
Does structured data affect attestation?+
It helps. Organization, Product, Article and FAQPage markup make entities and claims machine-readable and reinforce consistency, which supports the engine's judgement even though markup alone does not guarantee a citation.