An AEO audit checks whether AI engines can find your brand, read your pages, verify your facts, and cite you in the answers they write. It is not a score for its own sake, and it is not a list of tips. It is a check of the path from a question to an answer: can the engine reach the page, can it read the page, can it trust the facts on the page, and does the page answer the question the shopper asked?
Google says optimizing for generative AI search is optimizing for the search experience, and thus still SEO, and that a page must be indexed and eligible to be shown with a snippet to appear in its AI features. An AEO audit starts from that gate, and then checks the facts, the structure, and the answers that sit behind it.
Where it goes wrong
Audits fail when they measure the wrong thing, or when they measure the right thing and then stop.
- The audit only checks the homepage. A homepage does not carry the product facts, the prices, or the stock. An audit that never opens a product page, a category page, or a guide will miss the pages that carry the sale.
- The audit only checks the score. A score with no page attached is a number with no work attached. An audit that reports a score, and never names the page, the fact, and the fix, is a report, not an audit.
- The audit checks a tool, not the store. A tool that reports a low citation rate is reporting what the engines do with the pages they can read. An audit that blames the tool, and never opens the page, will not find the cause.
- The audit treats a friction point as an abort. A page that is slow, or hard to read, is a friction point. A real abort needs an agent genuinely attempting a purchase and failing. An audit that calls every friction point an abort will overstate the problem and lose the team's trust.
- The audit is done once and never repeated. A store changes every week: prices move, stock changes, products are added and removed. An audit that is never repeated goes stale, and the team acts on old facts.
The steps
Step 1. Check whether the engines can reach the site. The marketer and the developer check robots.txt together. Which crawlers are allowed, which are blocked, and which are slowed? Check the file at the site root, and confirm the rules match what the team meant. RFC 9309 says plainly that robots.txt rules are not a form of access authorization, so a block in robots.txt is a request, not a lock, and enforcement happens at the edge.
Step 2. Check whether the pages are indexed and eligible. The marketer checks Search Console for the pages that matter: are they indexed, and are they eligible to be shown with a snippet? Google says a page must be indexed and eligible to be shown with a snippet to appear in its AI features. A page that is not indexed cannot be cited, however good its content is.
Step 3. Check whether the facts are on the page, in words and in code. The developer opens a product page and compares the words with the code. Does the page state the price, the availability, and the shipping terms in plain words? Does the Product block state the same price, the same availability, and the same shipping terms in JSON-LD? Google says structured data is not required for generative AI search, and no special markup is required, but the facts still need to be readable, and the markup is how a system verifies them without guessing.
Step 4. Check whether the pages answer real questions. The marketer reads the product pages, the category pages, and the guides against a list of real buyer questions. Does the page answer which product suits a small kitchen, which shoe fits a wide foot, or which food suits a sensitive stomach? A page that only lists a name and a price does not answer a question, and an engine that cannot find an answer will cite a page that does.
Step 5. Check what the engines actually say about the brand. The marketer asks the engines the brand's own questions, and reads the answers. Does the answer name the brand? Does it state the facts correctly? Does it cite the brand's own page, or does it cite a competitor, a review, or a comparison? A wrong fact in an answer is a finding, and the fix is on the page the answer should have cited, not in the engine.
Step 6. Check whether an agent can buy. The developer walks the purchase path the way an agent would: can the agent find the product, verify the price and the stock, and reach checkout? This is the ASO check, and it is part of an AEO audit when the brand wants agents to buy, not just to cite. A page that is readable but not buyable will be cited and then abandoned at checkout.
Step 7. Write the fix list, page by page. The marketer writes the fix list with one line per page: the page, the fact that is wrong or missing, and the fix. No score without a page, no page without a fix, and no fix without a person who owns it. Then repeat the audit on a schedule, so a change in the score means a change in the store, not a change in the audit.
The machine-readable piece
An audit checks the machine readable facts first, because they are the easiest to check and the easiest to get wrong. This is the shape of a Product block an audit will read, using the schema.org vocabulary. Replace the placeholders with your own product facts, and only use values that match the visible page.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Product",
"name": "Example Trail Running Shoe",
"image": "https://[yourdomain.com]/images/trail-running-shoe.jpg",
"description": "A trail running shoe with a wide fit and a grippy sole for wet paths.",
"brand": {
"@type": "Brand",
"name": "Example Running Brand"
},
"offers": {
"@type": "Offer",
"price": "129.00",
"priceCurrency": "USD",
"availability": "https://schema.org/InStock"
}
}
</script>
Check the block after you publish it. Open the page, view the source, and confirm the name, the price, and the availability match what the shopper sees. Then run the page through the Rich Results Test and fix any error it reports. An audit that finds a clean block will move on to the harder checks, which are the words, the answers, and the purchase path. An audit that finds a block that disagrees with the page will stop there, because a system that reads a lie will not trust the next fact on the page.
What it looks like when it works
Say you run a running shoe brand, selling trail shoes, road shoes, and racing shoes, with a catalog of about one hundred products. The marketer runs an audit that checks robots.txt, indexing, product facts, real questions, engine answers, and the purchase path, and writes a fix list with one line per page.
The audit finds three things. The trail shoe page is indexed and readable, but it has no answer to the question which shoe fits a wide foot, so the engine cites a competitor. The road shoe page states the price in words but the Product block states a different price, so the engine does not trust either. The racing shoe page is readable but the checkout button is hidden behind a script the agent cannot reach, so the agent aborts at checkout. The marketer fixes the three pages, one line each, and the next audit shows the brand cited for trail shoes, the price trusted for road shoes, and the racing shoe buyable. The audit did not make the sale. It named the pages, and the pages did the work.
The fundamentals
These mechanics are shared across every engine and covered in full in their own guides:
- •The llms.txt format and a complete sample, see our guide to writing a good llms.txt file
- •The markup AI systems read to verify facts, see our product schema guide
- •The complete crawler rule reference, see our robots.txt guide for AI search
What this guide does not cover
This guide covers what an AEO audit actually checks: reach, indexing, facts, answers, engine answers, and the purchase path. It does not cover the full technical audit that checks every crawler and every page in detail, the transaction layer that lets an agent complete a purchase in full, or the per-engine playbooks for getting each assistant to find you. It does not cover how to run an audit tool, or how to price an audit. Those are separate jobs, and this guide is the map, not the whole trip.
Check your own store now
Run the free scan. One click, no signup, and it shows whether AI crawlers can read your store today: aizii.ai/aeo-scan
Run the free scanSources
- Google Search Central, Optimizing your website for generative AI features on Google Search (specific): https://developers.google.com/search/docs/fundamentals/ai-optimization-guide
- Google Search Central, Product structured data (specific): https://developers.google.com/search/docs/appearance/structured-data/product
- RFC 9309, Robots Exclusion Protocol (specific): https://www.rfc-editor.org/rfc/rfc9309.txt
- Schema.org, Product type (specific): https://schema.org/Product
- Schema.org, Offer type (specific): https://schema.org/Offer
Note: links marked specific point to the page that supports the claim in this guide. Links marked general docs point to the vendor's general docs section for that topic, not to a page that proves every step. Check the vendor docs for the current path before relying on a menu name or click path in this guide.
