Grade Your Content for AI Citations
Find out if your content is structured for AI search engines to cite. Check question headings, answer blocks, schema, source citations, and more.
Want us to handle everything?
Our managed SEO, AEO, and GEO services do the optimisation work for you. Get found where AI looks.
What makes content citation-ready?
Citation-ready content answers a real question directly, in a format a machine can quote. That means headings shaped like the questions customers ask, an answer in the first sentence rather than the last, schema markup that tells engines what the page is, and claims backed by sources.
None of it is exotic. It is mostly the discipline of writing for a reader in a hurry, applied to a reader that happens to be a machine. The grader shows you exactly where a page falls short.
What the grader measures, and why it uses fixed rules.
Enter a page address and the grader fetches that page once and scores it on five things AI engines tend to reward when they choose a source to quote. It uses fixed rules rather than an AI model, so the same page always gets the same score, and you can see exactly why it scored the way it did.
Question headings, worth a fifth of the overall score, measures what share of your headings are phrased as questions or start with words like what, how and why (about a third earns full marks). Answer blocks, worth a quarter, counts paragraphs of thirty to eighty words: long enough to answer something, short enough to quote. Schema, a fifth, looks at the JSON-LD types on the page and scores higher when there are several, especially FAQ or HowTo. Sources, worth fifteen per cent, counts outbound links and looks for figures and references to research. Structure, the final fifth, looks at length, headings, lists, a single main heading and the use of subheadings.
A page does not need full marks everywhere. A service page will rarely have question headings or outbound sources, and that is fine. The pages where question headings and answer blocks matter most are the ones meant to answer questions, such as guides, FAQs and comparison pages, so start with those, and aim for the areas that suit each page's job.
A few things to know about how it reads a page. The grader reads the page as it is first delivered, before any scripts run, so schema added by JavaScript after the page loads will not be seen. It looks at the whole page, including your navigation and footer, so footer and social links count as outbound links, and the sources score is a rough guide rather than a precise measure. And it works best with a page address: pasted text arrives without headings, paragraph tags or links, so it scores far lower than the same words on a real page.
If your answer blocks score low, look for long paragraphs that bury the answer, and split them so the first two or three sentences answer the question on their own. If question headings score low, rewrite a few headings as the questions your customers actually ask. If schema scores nothing, the schema generator on this site will give you a starting point, and the grader will pick it up on your next run. Then grade the page again: because the rules are fixed, any change in the score comes from your edit, not from the tool changing its mind.
The grade and the five scores appear straight away. Add your email to see the findings behind each score; we send you your score with a link back here, followed by a couple of follow-up emails you can unsubscribe from at any time. You can grade ten pages a day.
Grading one page tells you how quotable that page is. To see how AI engines treat your business as a whole, run the visibility checker or the free audit.
Making sense of your grade.
What does the grader actually score?
Five things AI engines look for when choosing what to cite: question-shaped headings, paragraphs short enough to quote as an answer, schema markup coverage, source citations and data-backed claims, and overall structure a machine can parse. You get a score per area and the findings behind it, not just a number.
What counts as a good score?
Treat anything below the top band as a to-do list rather than a fail. Most pages were written for human readers and Google, not for machines that quote, so middling first scores are the norm. The point is the gap list: fix the cheapest items first and re-grade.
Why do AI engines care about structure?
Because they quote rather than rank. An engine assembling an answer favours content it can lift cleanly: a question heading that matches the query, followed by a direct answer in the first sentence or two. Buried answers and clever headlines read fine to humans and poorly to machines.
How do I fix a low score?
Rework headings into the questions customers actually ask, open each section with the answer, add the missing schema, and cite your sources. If you would rather not do it by hand, the free audit shows where the same problems appear across your whole site.
One page graded.Now do the whole site.
The free audit reads your live site the way an AI engine does, then tells you whether ChatGPT, Perplexity and Google’s AI Overviews name you, or a competitor, when a buyer asks. Sixty seconds, no card.
