How to Audit Your Website for ChatGPT Shopping Visibility

What Is ChatGPT Shopping?

ChatGPT Shopping is one of the more interesting developments I've been tracking. OpenAI's product discovery feature lets users find and compare products directly inside ChatGPT conversations. Ask "What's the best espresso machine under $500?" and ChatGPT can display product cards with images, prices, reviews, and direct purchase links.

Unlike Google Shopping, which pulls from a merchant feed, ChatGPT Shopping relies on web crawling and structured data extraction. Your existing site is the source of truth. When I started adding e-commerce checks to my scanner, I realized how many sites fail at the basics here — the product data is either missing, malformed, or invisible to crawlers.

A Caveat Before We Start

I want to be upfront about something. I've scanned hundreds of sites and run a correlation study between structural readiness scores and actual LLM citations. The result was r=0.009 — essentially zero correlation. Structural readiness does not guarantee visibility.

What I believe these checks represent are prerequisites, not guarantees. A site that blocks AI crawlers or has no Product schema will almost certainly not appear. But fixing those issues does not mean you will appear either. Think of it as removing barriers rather than buying a ticket.

With that honest framing, here is what I check and what I consistently find broken.

The 7-Step ChatGPT Shopping Audit

Step 1: Check Robots.txt for AI Crawler Access

This is the single most common blocker I find. Roughly a third of e-commerce sites I scan either explicitly block AI crawlers or have broad wildcard rules that catch them by accident. Check your robots.txt for these user agents:

• OAI-SearchBot — used for ChatGPT search and shopping features
• ChatGPT-User — used when ChatGPT browses on behalf of a user
• GPTBot — used for training (optional, but allowing it seems to help)

# Example: Allow all AI crawlers
User-agent: OAI-SearchBot
Allow: /

User-agent: ChatGPT-User
Allow: /

User-agent: GPTBot
Allow: /

If you see Disallow: / for any of these agents, or a blanket User-agent: * with broad disallow rules, your products will not appear in ChatGPT Shopping. I've seen Shopify stores where a security plugin added blanket blocks without the owner knowing.

Step 2: Validate Product Schema Markup

When I scan e-commerce sites, the most common issue I find is not missing schema — it's incomplete schema. Many platforms generate a basic Product type automatically, but leave out fields that matter for shopping contexts. Here is what I consider the minimum viable Product schema:

{
  "@context": "https://schema.org",
  "@type": "Product",
  "name": "Aqualung Caldera Dry Suit",
  "description": "Professional-grade dry suit for cold water diving...",
  "image": "https://example.com/images/caldera-dry-suit.jpg",
  "brand": {
    "@type": "Brand",
    "name": "Aqualung"
  },
  "sku": "AQL-CLD-001",
  "gtin13": "3610391234567",
  "offers": {
    "@type": "Offer",
    "price": "1299.00",
    "priceCurrency": "EUR",
    "availability": "https://schema.org/InStock",
    "url": "https://example.com/products/caldera-dry-suit"
  },
  "aggregateRating": {
    "@type": "AggregateRating",
    "ratingValue": "4.7",
    "reviewCount": "89"
  }
}

The fields I see missing most often: brand, sku, and gtin13. Without these, a crawler has no reliable way to identify your product versus a competitor's listing of the same item.

Step 3: Verify Pricing and Availability Data

ChatGPT Shopping displays price and availability directly in product cards. I built a specific check for price mismatches in my scanner because the problem is so common. Here is what to verify:

• Every product page has an Offer with price and priceCurrency
• Availability uses valid schema.org values: InStock, OutOfStock, PreOrder
• Schema price matches the visible price on the page (mismatches hurt trust)
• Sale prices use priceValidUntil to indicate expiration

The mismatch issue deserves emphasis. I regularly find sites where the schema says one price and the page shows another — usually because a sale was applied in JavaScript after the static HTML rendered. From a crawler's perspective, the schema price is what it trusts.

Step 4: Test Review and Rating Aggregation

Product cards in ChatGPT Shopping show star ratings when available. In my experience, sites that include review data stand out more in shopping results. Check that:

• aggregateRating is present with ratingValue and reviewCount
• Individual Review schema is present (not just the aggregate)
• Reviews are crawlable (not loaded only via JavaScript after user interaction)

That last point is a pattern I see constantly. Many review widgets load via third-party JavaScript — Yotpo, Judge.me, Trustpilot embeds. The reviews exist for human visitors, but a crawler fetching the static HTML sees nothing.

Step 5: Audit Image Quality and Alt Text

ChatGPT Shopping displays product images in cards. This is one area where the requirements are straightforward:

• Primary product image is at least 800×800 pixels
• Image URL in schema matches a high-quality, non-watermarked image
• Alt text is descriptive (e.g., "Aqualung Caldera dry suit front view" not "product-123.jpg")
• Images load without JavaScript (server-rendered or static)

I check alt text coverage as one of my 26 audit checks. The typical e-commerce site I scan has about 40-60% of product images with meaningful alt text. The rest are either empty, auto-generated filenames, or generic strings like "product image."

Step 6: Check Page Load Performance for Crawlers

AI crawlers have timeout limits. This is where JavaScript-heavy sites run into the most trouble. My scanner compares what a simple HTTP fetch returns versus what a full browser render shows. The gap is often dramatic:

• Product data must be in the initial HTML response (not loaded by JavaScript)
• Page should respond within 3 seconds
• JSON-LD schema should be in the <head> or early in the <body>

I apply a JS dependency penalty in my scoring model: if the static HTML contains 50 words or fewer, the entire Machine Readability score gets halved. It sounds harsh, but if a crawler sees an empty page, nothing else matters. React-based storefronts without server-side rendering fail this check almost universally.

Step 7: Test with the ChatGPT Interface

The final step is a live test. This is where theory meets reality, and where you may discover that fixing all the structural issues still does not guarantee visibility:

1. Open ChatGPT and ask a product question relevant to your inventory
2. Try both generic queries ("best dry suit for cold water diving") and specific ("Aqualung Caldera dry suit price")
3. Check if your product appears in the shopping results
4. If not, compare your product page with a competitor that does appear — note what they have that you lack

I want to set expectations here. Even with perfect schema, open crawler access, and fast server-rendered pages, your products may not appear immediately. ChatGPT Shopping's selection logic is not transparent, and crawl frequency varies. What this audit does is remove the known barriers. The rest depends on factors neither of us fully controls.

ChatGPT Shopping Requirements vs Traditional SEO

I find this comparison helpful for e-commerce teams used to Google Shopping. The mental model is different enough that it catches people off guard:

Audit Checklist Summary

I use a version of this checklist in my own scanner. These are the items I check across 26 automated checks, distilled to what matters specifically for ChatGPT Shopping:

What This Audit Can and Cannot Do

I want to end with the same honesty I started with. This audit identifies structural barriers — things that will definitely prevent your products from appearing in ChatGPT Shopping if left unfixed. Blocked crawlers, missing schema, JS-only rendering: these are clear, fixable problems.

What this audit cannot predict is whether fixing them will make you visible. My research across 441 domains showed near-zero correlation between readiness scores and actual AI citations. The one real signal I found was content relevance — being topically aligned with the query matters far more than any structural checklist.

So treat this as removing known blockers. It is necessary work, but not sufficient on its own.

Next Steps

Start with our free AI Search Readiness audit to see where your site stands on these checks. Then work through this 7-step checklist to address what comes up. For ready-to-use JSON-LD examples, see our Schema.org Markup Guide for E-Commerce.