Why do some Shopify products show up in ChatGPT, and yours don't?

Since March 2026, every Shopify store is shoppable inside ChatGPT, Perplexity and Copilot. The assistants decide which products to recommend by reading product pages. Most stores have never been checked against what they actually read.

When someone asks ChatGPT for "a heavyweight cotton tee that won't shrink," it doesn't rank pages the way Google does. It reads product data: the title, the structured data, the identifiers, the reviews, whether the thing is in stock. If your pages don't say the boring details clearly, you don't exist in that conversation. A competitor with better-filled-in fields does.

Shopify's own dashboard will now give you a visibility score. What it won't tell you is which of your 340 products are invisible, why each one is being skipped, or what to change. That gap is what this tool covers.

What the audit checks

• Whether the AI crawlers (GPTBot, ClaudeBot, PerplexityBot) can reach your pages at all. CDNs block them by default more often than you'd think.
• Structured data on every product page: price, availability, ratings, GTIN/MPN identifiers.
• Whether your reviews are readable by crawlers or locked inside a JavaScript widget they never execute.
• Titles and descriptions that state material, size, fit and use, not just a product name and a vibe.
• FAQ content that answers the questions buyers actually type into assistants.
• The small plumbing: sitemap, llms.txt, feed consistency between your pages and your catalog.

What you get

A report that goes product by product: what's missing, what to write instead, and the corrected fields as a CSV you can re-import into Shopify in one step. No app to install, nothing to grant access to. It runs from your store URL.

What we're finding so far

Before opening this up, I've been calibrating the checks on 17 well-known Shopify stores. Early patterns, July 2026:

• Scores ranged from 16 to 75 out of 100, and size didn't predict the score. Some of the biggest brands did worst.
• Three household-name stores serve product pages with no machine-readable product data at all. Their custom storefronts render everything in JavaScript, which AI crawlers don't execute. A small, well-configured store can genuinely beat them on these surfaces.
• The most common single problem: review ratings locked inside a JavaScript widget, invisible in the structured data. The social proof exists; the machines just can't see it.

The full study, with methods and store-by-store data, will be published here.

Questions you'd reasonably ask

Is this a Shopify app I have to install?

No. It works from your public store URL. You don't install anything or grant any access. An app version with one-click fixes is planned, but only once the audit itself has proven useful.

What does it cost?

The first audit is free while I calibrate the tool. The full version will be a flat monthly price per store, in the range a normal store budgets for one app. Not an enterprise "contact us."

Does AI shopping traffic even matter yet?

It's early, and I won't pretend otherwise. Assistants send a fraction of a percent of Shopify traffic overall today. But stores that show up well already report 5% or more of orders from it, and buyers arriving from an assistant convert at several times the rate of search visitors, because they arrive pre-convinced. The point of fixing this now is that it's cheap to be early and expensive to be late.

Who's building this?

I'm Jake. I build small, focused tools, and I built this one after watching merchants in the Shopify community test prompts by hand to figure out whether ChatGPT knew their store existed. That seemed like a job for software.