How the AI shopping readiness checker works
The checker fetches your page the way an AI crawler does. It reads the raw HTML and runs no JavaScript, because many AI crawlers don't. It then reads your robots.txt and llms.txt files.
It scores fifteen checks out of 100. The product schema, price, identifiers, shipping, returns and readable text carry the most weight. Those are what AI shopping tools filter and compare on. In Adobe's April 2026 benchmark, AI could read only 66% of an average US retail product page.1
It also checks nine crawlers by name. Four feed AI search and shopping answers: OAI-SearchBot, Claude-SearchBot, PerplexityBot and Googlebot. Only those four count towards the score. Training crawlers, such as GPTBot, are shown so you can decide. OpenAI's own documentation explains that its search and training crawlers are set separately.2 The reasons behind each check are in the AI for ecommerce guide.
Questions people ask
What score should a product page get?
Aim for 85 or more. Product schema with price, stock, an identifier, shipping and returns, readable text and open AI search crawlers will get you there on most platforms.
My store is on Shopify. Why is the score low?
Many themes skip shipping and returns markup, and some leave out the GTIN. Fill those fields in your product data, then check whether your theme outputs them. Apps can add the missing markup.
Should I block GPTBot and other training crawlers?
That's your call. Blocking them keeps your content out of model training. It doesn't stop ChatGPT search, which uses OAI-SearchBot, as long as that crawler stays allowed.
Does the checker store the pages I test?
No. It fetches the page once to score it and keeps nothing. Results may be cached for an hour so repeat checks are faster.