Citation recall
Citation recall is the share of the statements in an AI answer that the sources it cites fully back up.
Citation recall is the share of the statements in an AI answer that its cited sources fully back up. If an answer makes ten claims and the pages it links to support eight of them, its citation recall is 80%.
It matters because a link next to a sentence looks like proof. Few readers click through to check, so a claim with a citation that doesn't support it carries the same weight as one that does.
Where the term comes from
The term comes from research on AI search. A 2023 study, Evaluating Verifiability in Generative Search Engines, checked answers from four AI search engines by hand and scored them on two things. Citation recall asks whether everything the answer says is backed by a citation. Citation precision asks whether each citation actually supports the sentence it's attached to.
A trustworthy answer needs both. High recall means nothing is said without a source. High precision means the sources say what the answer claims they say.
Why it matters for your brand
For example, say an answer reads "Fernhill Coffee roasts in small batches and only delivers in London," and links to Fernhill's About page. The first half is on that page. The second half isn't, and it's wrong. A reader sees one neat citation and believes both halves.
So when you read answers about yourself, don't stop at whether your page was cited. Check that the sentence next to the link says what your page says. An answer can quote your site and still get you wrong.
In AnswerOrbit
The closest thing is Claims. We check every claim an answer makes about you against your facts, and when a claim contradicts one, we quote the sentence it came from. Inaccuracy drivers shows the pages behind a wrong claim. If an answer said something wrong without using any page, you'll see the platform's own knowledge listed instead.