Evergreen AI
Get started
Free AEO checker, no signup, 15 checks in seconds

Will AI quote my page?

Paste any page address. We check the 15 things that decide whether an answer engine can lift a passage out of it and stand behind the quote, then tell you which of them are failing and what to change.

Works on any public page. We never ask for access to your site.

Scored for the way answers are assembled by

  • ChatGPT
  • Claude
  • Perplexity
  • Gemini
  • Copilot

What answer engine optimization actually is

Search used to end with a list of links. Increasingly it ends with an answer, assembled by ChatGPT, Perplexity, Claude, Copilot or Google's AI Overviews, with a handful of sources cited underneath it. Being one of those sources is a different job from ranking.

The practical difference is what gets used. A search engine ranks pages. An answer engine quotes passages. To quote you it has to find the passage that answers the question, satisfy itself that the passage stands up on its own, and decide your page is worth standing behind. A page can rank well and still fail all three, which is why an SEO audit can come back clean while the thing your customers now ask never mentions you.

That is what this checker measures. It does not guess at rankings and it does not grade your writing. It looks at the mechanical properties that decide whether a machine can use your page, and tells you which of them are failing.

Is this an answer engine optimization tool or a generative engine optimization tool?

Both, and there is no difference worth arguing about. Answer engine optimization (AEO) and generative engine optimization (GEO) name the same work: making a page that a generative system can find, read, lift a passage from and cite. AI SEO is a third label for it. Nobody has won that argument yet, and the checks below would be identical under any of the three names.

If you came here looking for a GEO checker, this is it. We use AEO throughout because it says what the machine is doing, and because "answer engine" describes the output a reader actually sees. Where the two schools differ in practice is emphasis: GEO writing tends to talk about being quotable at length by a model that is generating, and AEO about being extractable at the passage level. The measurements are the same measurements.

What the label does not change: the three groups below. Structure and extractability, Trust signals, Fundamentals. 15 checks in total, because none of the naming schools has a check the others do not.

Have you optimised for featured snippets? You have already done most of this

This is the least appreciated thing about AEO, and it is worth saying plainly rather than selling the work as new.

Featured snippets, the answer box Google has shown above its results for years, are won by a specific and well documented page shape. Ask the question in a heading. Answer it immediately underneath, in a short self-contained block. Use a list when the answer is a sequence, and a table when it is a comparison. Google publishes its own guidance on this, and an entire generation of SEO work went into it.

Read that description again and it is nearly a specification for the structure group in this checker. That is not a coincidence. A featured snippet is an extractive answer, and so is a citation in ChatGPT or Perplexity: in both cases a machine has to decide where an answer starts and stops, take it out of your page, and present it with nothing else around it. The mechanism is the same even though the surface changed.

The bridge is not perfect, and the differences matter:

  • A snippet quotes one passage from one page. An answer engine synthesises several sources, so a page that supplies one clean, attributable fact can be cited alongside four others rather than having to beat them.
  • Snippet work optimises for one ranking position. AEO has no position to win, which removes the winner-takes-all part and makes the trust signals matter more: a model choosing between three pages that say the same thing leans on authorship, dates and citations.
  • Snippets are a Google feature with Google rules. Answer engines are several vendors with no published rules at all, so the honest strategy is the structural one that works everywhere rather than tuning for one of them.

None of that changes the practical advice. If your pages already win snippets, run one of them through this checker and you will most likely find the structure group is nearly clean and the trust group is where the work is.

What we check, and why

15 checks in three groups, every one measured from your page rather than estimated. Structure decides whether an answer can be lifted out. Trust decides whether the thing lifted is worth quoting.

Structure and extractability

An answer engine does not quote a page, it quotes a passage, so it has to find where an answer starts and stops. We check whether your headings form a clean outline, whether any of them are phrased as a real question, whether a short self-contained answer sits directly beneath each one, and whether you use lists or tables anywhere. A page can be thorough and still be unquotable when the answer is four paragraphs down.

  • Headings are read from the article itself, with navigation, headers, footers and sidebars stripped out first, so a footer column heading is not mistaken for your content.
  • A question heading has to contain an actual question mark. "Pricing" is a topic, "How much does it cost?" is a question, and only the second one matches how people prompt.
  • An answer of roughly 10 to 120 words directly under the heading is what gets lifted. A list or a short table counts, and is often better than a paragraph.

Trust signals

AI systems weigh whether a source looks accountable. We check for structured data, a named author, a stated published or updated date, outbound links to the sources behind your claims, and an FAQ block. None of these are hard to add, and together they are a large part of why one page gets cited over another that says the same thing.

  • Structured data is read from JSON-LD, the format described at schema.org and the one every major consumer of it prefers.
  • A date only counts if a machine can find it, which usually means a dateModified in your schema as well as a line a reader can see.
  • Outbound links are counted inside the article, and nofollowed ones do not count, because a citation you refuse to endorse is not a citation.

Fundamentals

Two checks, not the usual seven. A title and a meta description are what a machine uses to decide what this page is about before it reads a word of the body, so they earn their place here. Title length bands, canonical tags, alt-text coverage and redirect hops do not: they are worth fixing, they are simply not why nobody quotes you, and bundling them in was what made an AEO result read like a generic SEO audit.

Crawler access and JavaScript rendering are not checked here. They are properties of a site rather than of a page, they are answered yes or no rather than scored, and averaging a locked door into a gradient is what stopped the old combined score meaning anything. If you are not certain an AI can reach your site at all, start with the AI Visibility Checker and come back here afterwards. Vocabulary for the structured data below is defined at schema.org, and Google documents which types it consumes.

How this works, and what it cannot see

Every limit of the method, stated by us rather than discovered by you.

What this tool reads

One URL, fetched once, as raw HTML. Every finding is measured from that single response. Nothing is estimated, nothing is inferred from your domain or your industry, and nothing here looks at your robots.txt or your crawler rules: whether a machine is allowed to reach your site at all is a different question with a different answer, and it has its own tool.

How structured data is read

Structured data is read from JSON-LD only. Microdata and RDFa, which many Shopify and WooCommerce themes emit and which Google does read, are not detected here, so a page using them can be reported as having none. That is a limit of our reading, not a verdict on your page.

Where the AI is used, and where it is not

One part of the result, the passage breakdown, uses a small AI model to judge which sections of your page could stand on their own if quoted. It is the only AI in the tool. Every check and the whole score are deterministic, so if that step is skipped or unavailable your result and your number are unchanged.

What this tool cannot see

The structure and trust checks read the article rather than the whole document, and they find the article by stripping semantic navigation, header, footer and sidebar elements. That works where the markup is semantic. On a site that builds its footer out of unmarked divs, some of that footer still counts as content, which usually shows up as a page looking better supplied with links and headings than it really is.

How the score is calculated

A weighted pass rate over whatever we were able to check, so a check we could not run leaves the calculation rather than counting as a pass or a failure. A warning earns half credit, and seven checks are warn-only by design, which is why a page that passes everything failable and warns on everything advisory lands in the middle band rather than the top one. If we could not read your page well enough to check anything substantial, you get no score rather than a zero.

What we keep afterwards

We keep the address, each check outcome, the score, a few page-level measurements and the passage summaries, so a shared result link keeps working. We do not keep a copy of your page, its HTML or its text. Checking the same address twice within a day returns the stored result instead of fetching you again. Last updated 24 August 2026.

Questions people ask about AEO

What is answer engine optimization?

Answer engine optimization, or AEO, is the practice of structuring a page so AI systems can find it, understand it and quote it. You will also see the same thing called generative engine optimization (GEO), or simply AI SEO. The terms are used interchangeably and none of them has settled yet. The underlying question does not change with the label: when someone asks an AI a question your page answers, can the AI actually use your page?

What counts as a question heading?

A heading containing a literal question mark, phrased the way somebody would ask it. "Pricing" is a topic and scores nothing. "How much does an AEO audit cost?" is a question and matches how people prompt an assistant. The check is deliberately strict about the question mark, because a strict rule means a pass is real evidence rather than a guess about intent. Two or three genuine question headings do more for a page than a dozen noun-phrase ones.

How close does the answer have to be to the question?

Directly underneath, in the first block after the heading, and roughly 10 to 120 words long. That is the window an answer engine can lift without dragging in the paragraph before or the caveat after. Below about ten words there is no answer to take. Above about a hundred and twenty it stops being a passage and starts being a section. A short list or a small table counts, and often reads better than prose.

How do I get cited by ChatGPT?

Make one passage worth lifting. Give the specific question a heading of its own, phrased as the question, and answer it in the first paragraph underneath in a way that makes sense with nothing else around it. Then make the page look accountable: a named author, a date, structured data, and links out to whatever you are relying on. All of that assumes ChatGPT can already reach and read your site, which is worth confirming separately.

How do I rank in ChatGPT and other AI search tools?

There is no ranking to win in the sense SEO means it. An assistant assembles an answer and cites a handful of sources under it, so the goal is being one of those sources. In practice that means being findable by ordinary search, being quotable at the passage level, and looking like something worth standing behind. The first is old work, the second and third are what this checker measures.

How is this different from a normal SEO audit?

A traditional SEO audit asks whether Google can crawl and rank you. This asks whether an AI can quote you, which is a different problem. An answer engine has to lift a specific passage out of your page and stand behind it, so it cares about whether your answers are self-contained, whether your headings match the questions people actually ask, and whether the page looks accountable. A page can rank on page one and fail every one of those.

Does AEO replace SEO?

No, and treating it as a replacement is the most expensive mistake in this area right now. Answer engines are largely built on top of the same signals: they still need to crawl you, still need to find you relevant, and mostly cite pages that already rank. AEO is what you add once the SEO fundamentals hold. Almost everything this tool checks helps both, which is the honest reason to do it.

Is the AEO Page Checker free?

Checking one page is free, with no signup and no email. There is a rate limit per visitor, which is there to stop the tool being scraped rather than to push you towards an account. If you want every page on your site scored, tracked over time and rewritten rather than only measured, that is what Evergreen AI does and what an account is for.

Do you store my page?

We store the address you checked, what each of the checks concluded, the score, and a handful of page-level measurements such as your title, meta description and heading counts, so a shared result link keeps working. We do not keep a copy of your page, its HTML or its text.

One page checked is a sample, not a strategy

This checker tells you where a single page stands today. Evergreen AI does the rest of it continuously: every page scored, competitors watched, new content written to be quoted, and existing pages re-optimised as the answer engines change their minds.

Get started free