What is the AEO Readiness Audit?
| What it answers | Score any URL out of 100 on answer-engine extraction — direct answers, question headings and schema. |
|---|---|
| How the answer is produced | Answer Engine Optimization is a narrow question with a specific answer: when a system like Google AI Overviews, Perplexity or ChatGPT search is assembling a reply, can it lift a short, self-contained passage from your page and present it as the answer? |
| What you need to enter | Paste the URL you actually want quoted — a reference page or guide that answers a specific question, not a homepage that answers none. |
| Where it stops being reliable | It scores one URL at a time and cannot see pages that require authentication. |
| Cost and sign-up | Free, runs in your browser, no account and no stored inputs. |
How is the AEO score calculated?
Answer Engine Optimization is a narrow question with a specific answer: when a system like Google AI Overviews, Perplexity or ChatGPT search is assembling a reply, can it lift a short, self-contained passage from your page and present it as the answer? That is a different question from whether Google ranks you, and a different question again from whether an AI crawler is allowed to fetch you at all. This audit isolates it, so the score is not diluted by checks that belong to other disciplines.
Seven checks carry the weight here. Server rendering comes first, because a passage that only exists after hydration cannot be extracted by a crawler that never runs JavaScript. Canonical clarity follows, since an engine that cannot decide which URL is authoritative will often quote a duplicate. The remaining five are editorial: a direct answer in the opening paragraph, a takeaways block, question-shaped headings, facts presented as lists or tables rather than prose, and FAQ or HowTo schema that labels the answer explicitly.
Each check is scored pass, partial or fail from what was actually found in the HTML — word counts, canonical values, heading patterns, schema types — and weighted by how often it is the real blocker rather than how often it appears in checklists. The direct answer block and server rendering dominate, because a page failing either cannot be rescued by anything below it.
The remediation list is sorted by residual weight, not checklist order, so the first item is always the largest single available gain. In practice, most sites recover twenty to thirty points from the top two items alone, and both are usually editorial work that needs no engineering ticket.
How do you use the AEO Readiness Audit?
- 1.Paste the URL you actually want quoted — a reference page or guide that answers a specific question, not a homepage that answers none.
- 2.Read the evidence line under each check. It reports what was found in the raw HTML, so you can verify every verdict yourself with view-source.
- 3.If the page is behind a login or blocks automated requests, use the self-assessment checklist tab. A page a scanner cannot reach is a page an answer engine cannot reach either.
- 4.Work the fix list from the top, re-running the audit after each change, and run the GEO audit separately to check whether crawlers are permitted and whether you are credited by name.
What can this tool not tell you?
- It scores one URL at a time and cannot see pages that require authentication.
- It measures whether your page is extractable, not whether it is currently being quoted. Only prompting the engines or reading referral logs answers that.
- It scores structure, not authority. A perfectly formatted page on a domain nobody references still loses to a weaker page from an established source.
- Extraction behaviour changes with each model generation, so treat the weights as current practice rather than permanent rules.
Why extraction is where most pages lose?
Across audits, the technical layer is usually fine. Modern frameworks server-render by default, canonicals are generated automatically, and sitemaps produced by an untouched plugin still work. Teams arrive expecting a plumbing problem and instead find that crawlers can read every word of a page that contains nothing shaped like an answer.
That is a writing problem wearing a technical costume, and it is why the AEO score exists on its own. The fix is not a ticket for an engineer; it is rewriting the first two hundred words of the pages that matter, which is cheap in hours and expensive in ownership because no one team is responsible for it.
The resolution is not to abandon persuasive copy but to put one honest, self-contained paragraph in front of it. Readers who already know they are in the right place skim past. Readers who do not get their answer immediately and trust you more for it. The engine gets a block with a clean beginning and end that it can quote without editing.
Everything else on this list — takeaways, question headings, tables, schema — makes that paragraph easier to find and label. None of it substitutes for having written it.
What do worked examples look like?
A documentation page scoring 45
Server-rendered, canonical clean, schema present — and every article opened with two paragraphs about the product before defining the term in the title. Moving a forty-word definition to the top, unchanged in wording, took the same page to 78 on the next scan.
A comparison page scoring 62
The answer existed but was buried in a paragraph of hedged prose. Converting the six differentiators into a table and adding question-form H2s recovered nineteen points without adding a single new claim.
What do people ask most about this tool?
How is this different from the GEO audit?
AEO asks whether there is a passage worth lifting. GEO asks whether a generative crawler is allowed to reach the page and has a named, dated source to credit. Both matter, they fail for different reasons, and fixing one does nothing for the other — which is why they are scored separately.
What score should I aim for?
Sixty-five is the point where engines can reliably lift your wording. Above eighty-five the remaining constraint is editorial quality and reputation, not markup, and chasing a hundred is rarely the best use of an afternoon.
Why does my well-written page score badly?
Almost always because it opens with persuasion instead of an answer. Conversion copy withholds specifics to pull the reader down the page; an engine reads that opening, finds nothing repeatable, and quotes the competitor who stated the fact in sentence one.
Does FAQ schema alone get me quoted?
No. Schema labels an answer that already exists in the text. If there is no clean answer in the prose, marking up an empty one changes nothing. Write the paragraph first, then add the markup.
How long should the answer block be?
Forty to seventy words, self-contained, no pronouns pointing at earlier sentences, and no dependency on the heading above it. That is roughly the length engines lift intact.
How often should I re-run it?
After any template change, and quarterly otherwise. Editorial checks drift silently as pages are edited by people who never saw the original brief.
Which related tools should you try next?
Written and reviewed by Jim Vernon, Editor, AI Intelligence International. Published by AI Answer Engine, a service of AI Intelligence International, and checked against our editorial standards.