A 30-Minute ChatGPT Readiness Checklist for Shopify Stores

Minute 0-5: choose a comparison-heavy product

Pick one product with real buyer constraints. Open the canonical page and identify the selected variant, price, currency, stock, key attributes, package contents, and exclusions.

Minute 5-10: review Shopify channel state

In Shopify admin, review the product’s publication status and the Agentic or connected sales-channel settings available to the store. Confirm the product is eligible before assuming content is the problem.

Shopify Catalog can make eligible product data available to connected AI channels, but does not guarantee appearance.

Minute 10-15: read the page like a shopping researcher

Write one prompt:

Recommend a 65W travel charger under $80 with two USB-C ports,
a foldable plug, and support for 100-240V input.

Underline every required fact. Confirm the page states each one as text. Add the limitation that total output may be shared across ports.

Minute 15-20: brief the technical owner

Ask for a check of:

• OAI-SearchBot and ChatGPT-User policy
• public status and final URL
• canonical
• Product and Offer JSON-LD
• selected variant identity
• page/feed price and stock agreement

OpenAI documents different user agents for different purposes. Access alone does not prove indexing or recommendation.

See the OpenAI crawler guide.

Minute 20-30: run and record one observation

Use a clean ChatGPT session with Shopping Research or the relevant shopping-capable mode when available. Save:

• exact prompt
• clarifying questions
• products shown
• source URLs
• factual errors
• date, market, and mode

If the store is not shown, compare the attributes of the shown products. If the test fails, mark it unavailable rather than a non-mention.

The output

At the end, create one task:

DockPro page lacks input voltage and included-accessory facts.
Add visible specifications and package contents, then ask the developer
to reconcile the same SKU in JSON-LD and connected product data.

That is a useful readiness result. “ChatGPT score 64” is not.