The Quattr Method
R3 · Referenced, is your site the source AI answers use?
Citation share and mentions, never conflated with AI Overviews.
An assistant is answering a question about your category right now. Underneath the answer sits a short list of sources. That list is the whole prize, and most of the names on it belong to review sites, forums and publications that will never sell against you.
R3 · Referenced is the third of the Quattr Method's five questions: are you the source answers use, cited and mentioned? A citation means an AI answer named one of your pages as a source, counted across the questions you track. A mention means it named your brand without linking anything.
On answer surfaces the answer takes the click that used to be yours. Being the source behind it is the part still available. That makes this the question where AI search stops being a metaphor and turns into a scoreboard. Assistants write those answers either way. The only thing in play is whether they write them out of you.
A citation is the answer showing its work
The number here is citation share, with mentions reported beside it and never blended into it.
A citation puts your page in the source list, doing the answering. A mention means the model knows you exist, on grounds it did not disclose. The ordering holds. Citations above mentions, every time.
Engines keep separate ledgers too. ChatGPT, Perplexity, Gemini and AI Mode, Google's conversational search surface, all cite differently. A blended rate can hold perfectly flat while you trade one engine's wins for another's losses. Per-engine cuts come first. The aggregate is a summary of them.
The article on monitoring AI citations without a dashboard covers how to watch this week to week, and the ai-citation-monitoring skill runs it engine by engine, a packaged analysis you run by name in your assistant.
The workflow that does this: ChatGPT citations by cited page →
Referenced is not AI Overviews
The most common corruption of this question is conflation.
An AI Overview is the AI-written block Google shows on top of a classic results page, and you measure it as show-rate and inclusion. Answer-engine citations are an assistant crediting the sources it used, and you measure those as citation share. Different surfaces, different metrics. The method never sums or swaps them.
The conflation is not harmless bookkeeping. AI Overview inclusion responds to work on the results page. Citation share responds to whether your pages can be fetched and whether their passages stand up alone. Blend the two and you will eventually fund the wrong work with a straight face.
The article on AI Overviews and answer-engine citations takes the distinction apart. Here it is enough that a Referenced number mixing the two measures neither.
See also: why AI Overviews are not answer-engine citations →
Most of the inventory is not yours
A large share of what answers cite is third-party content you do not own. Review sites, comparison pages, forums and publications do a lot of the answering about your category, and none of that inventory is yours to rank.
That reframes who you are competing with. A search competitor and a business competitor are different animals. An aggregator taking the answer slot for your buyers' questions competes with you here even though it will never sell against you. The response is to answer those questions on your own domain, well enough to be fetched.
It also changes the work. Instead of only chasing rankings for pages you own, you ask which third-party surfaces answer for your buyers, whether what they say about you is still accurate, and whether your best material gives an engine something better to draw from. This is a question you share with the whole web that talks about you.
Ask it yourself
Which of our pages do AI answers cite most, and which prompts cite competitors but never us?
Which pages, and which prompts?
Two lenses, plus the gap between them.
One check shows which of your pages earn citations, how often, and where they sit in the source list. A second flips the lens to which prompts cite you most. A third names the prompts where you never appear. Together they turn a vague question, how visible are we, into three piles of work.
- Pages already earning citations, which you protect.
- Prompts where a competitor's page does the answering, which you contest.
- Prompts nobody has written a fetchable page for yet, which you write.
Referenced is gated from below
Three kinds of work name Referenced as their primary target: net-new content, authority and off-site, and AI-answer visibility. All three sit behind the same physical constraint. You can't be cited if you were never crawled, which is why every citation diagnosis eventually drops to R1 · Reachable, the question of whether search engines and AI bots can fetch, render, and index your site.
This is the question most vendors sell first, because citation dashboards demo beautifully. The method earns it from below instead. That is a slower thing to say in a demo and a much faster thing to prove afterwards.