AEO Grader for any product page.
Before an answer engine can recommend your product, it has to read the page. The AEO Grader scores any URL 0–100 on the four structural signals that decide whether ChatGPT, Perplexity, or Google AI Overviews can cite it — and ranks the fixes by severity.
Rate limit: 1 run per 30 seconds per IP. The grader reads only public data.
What the four buckets check
Every finding maps to a documented fix. The weights reflect what moves citations in observed audit data, not a marketing scheme.
| Bucket · weight | What it checks | Fix guide |
|---|---|---|
| Structured data · 30 | Product JSON-LD presence and completeness — name, brand, SKU, GTIN, offers, availability, aggregateRating. Engines use it to resolve identity and price. | Schema for AI search |
| Content depth · 30 | Description length and answer coverage: does the page answer material, fit, compatibility, and care questions in quotable prose, with FAQ content where it helps. | What is AEO |
| Crawler access · 30 | robots.txt admittance for AI crawlers, canonical hygiene, and meta directives. A blocked GPTBot zeroes out everything else on the page. | Rank products in ChatGPT |
| llms.txt · 10 | A curated crawl map at the domain root pointing AI crawlers at the pages that describe the brand and flagship catalog. | llms.txt for Shopify |
How to read the score
Composite bands: 80–100 means the page is structurally ready and remaining gains come from content and review signal; 60–79 means one or two buckets are dragging, usually thin schema or a robots.txt drift; 40–59 means the page needs a planned fix cycle; below 40 means foundational work has not started and the page is effectively invisible to answer engines regardless of how well it ranks on classic Google.
Two cautions. First, the score measures readiness, not citations — a 90 on a product nobody queries still gets zero citations, and live testing belongs to the ChatGPT checker. Second, AEO practice is still forming; the grader weighs the signals with the strongest observed correlation to citation behavior as of mid-2026 and the weights will move as engines do. Definitions for every term in the report live in the glossary, starting with AI visibility score.
Most teams use the grader as a quarterly snapshot: run it on the best-seller, archive the number, fix the top three findings, re-run. The catalog-wide version of that loop — every SKU, every week, with the fix shipped as a diff — is the eCommerce Insights product, and the per-page methodology is identical, so the free grade you run today is comparable to the tracked score later.
One grade is a snapshot. The trend is the story.
eCommerce Insights runs the AEO checks weekly on every PDP, scores citation and agent-readability separately, and sends a Monday digest with the catalog delta and the week's biggest movers.
Ask AI about the AEO Grader
Have your favorite AI engine summarize this tool for your use case.
Frequently asked questions
What is an AEO score?
Does the AEO Grader query AI engines directly?
What URL should I grade first?
When should I use the JS-rendering option?
How often should I re-run the grade?
Is the AEO Grader a generative engine optimization tool?
One number for the board. Ten fixes for the team.
The AEO Grader turns "how are we doing on AI search?" into an answerable question.