The 30-Minute Shopify AEO and GEO Checklist

What you will verify

This checklist establishes a compact baseline for one commercially important product. It does not score every page and it does not claim to reproduce every AI answer. It verifies whether the product has a stable, public trail from URL to visible facts to machine-readable facts, then records a small buyer-question sample.

Create a working note with five columns: check, URL or prompt, observed result, timestamp, and change required.

Step 1: open the product as a public visitor

Select one product that represents the store. Open its canonical URL in a private browser window. Confirm that a visitor can see the name, selected variant, price, currency, availability, brand, primary image, use case, and important compatibility or ingredient constraints.

If critical information exists only after opening a tab, downloading a PDF, choosing an unavailable variant, or signing in, record that limitation.

Step 2: inspect redirects and status

Run:

curl -sSIL https://example.com/products/example-product

Expected result:

HTTP/2 200
content-type: text/html

One normal redirect from an alternate hostname can be acceptable. Record redirect loops, password pages, bot challenges, unexpected locales, and a final URL different from the canonical shopping destination.

Step 3: inspect the canonical URL

Open page source and find:

<link rel="canonical" href="https://example.com/products/example-product">

The canonical should identify the public product destination, not a search page, collection-filter URL, preview URL, or unrelated variant.

Step 4: compare Product JSON-LD with the page

Find every application/ld+json block and identify objects with @type: Product. Compare the name, URL, SKU, brand, image, offers, price, currency, and availability with the visible product and selected variant.

Record contradictions separately from missing optional fields. A valid JSON document can still describe the wrong variant or stale inventory.

Step 5: inspect discovery files

Request:

curl -sS https://example.com/robots.txt
curl -sS https://example.com/sitemap.xml
curl -sS https://example.com/agents.md
curl -sS https://example.com/llms.txt

Do not grade a file by presence alone. Verify that referenced product, collection, policy, shipping, returns, and contact pages are canonical and publicly reachable.

Step 6: define five buyer questions

Create one question for each intent:

Discovery: Which specialist stores sell [category] for [use case]?
Best for: What is the best [product] for [constraint]?
Compatibility: Where can I buy [product] compatible with [model]?
Comparison: Which store has the clearest [attribute] information?
Purchase: Where can I buy [product] with [shipping or availability constraint]?

Use the same wording for the baseline and later comparison.

Step 7: record evidence, not just a yes or no

For each completed result, save the platform, prompt, timestamp, visible answer, source URLs, matched storefront domain, brand string, and named competitors. Mark provider errors as unavailable rather than “not mentioned.”

Expected outcome

At the end of 30 minutes, you should have one prioritized storefront inconsistency and five reusable buyer questions. You should not have a promise that one repair will create an AI recommendation.

How to make the first change

Choose the smallest high-confidence repair: a broken canonical, contradictory availability, missing visible compatibility fact, or stale structured identifier. Publish it, request the public page again, and save the corrected output.

Rollback

This audit is read-only. If the later repair requires Theme code, duplicate the current Theme first and document the changed file and lines. Revert the single change if visible product behavior, structured data, or checkout information becomes inconsistent.