← Back to blog
json-ldstructured-dataai-commerce
March 21, 2026 3 min read

JSON-LD for AI Shoppers: The Structured Data That Makes or Breaks Your Visibility

AI shopping agents don't read your marketing copy. They read your JSON-LD. Here's what they look for, what most stores get wrong, and how to fix it.

json-ldstructured-dataai-commerce

When ChatGPT Shopping recommends a product, it doesn’t browse your website like a human. It queries structured data — specifically, the JSON-LD markup embedded in your product pages. If your JSON-LD is missing, incomplete, or poorly structured, AI shopping agents simply can’t see your products.

What is JSON-LD?

JSON-LD (JavaScript Object Notation for Linked Data) is a way to embed structured data directly into your web pages. For ecommerce, the key schema type is Product, which describes everything an AI agent needs to know: name, brand, price, availability, images, reviews, and attributes.

Here’s what a well-structured product JSON-LD looks like:

{
  "@context": "https://schema.org",
  "@type": "Product",
  "name": "Women's Lightweight Running Shoe - Wide Width (D)",
  "brand": { "@type": "Brand", "name": "Acme Athletics" },
  "description": "Lightweight neutral cushion running shoe...",
  "sku": "AA-RUN-W-WIDE-001",
  "gtin13": "0123456789012",
  "image": "https://example.com/shoe.jpg",
  "offers": {
    "@type": "Offer",
    "price": "129.00",
    "priceCurrency": "USD",
    "availability": "https://schema.org/InStock",
    "itemCondition": "https://schema.org/NewCondition"
  }
}

What AI agents actually look for

Different AI shopping platforms prioritize different parts of your structured data:

ChatGPT Shopping

Pulls primarily from Google Shopping data, which itself relies heavily on your Google Merchant Center feed and product page markup. Strong titles with [Brand] + [Type] + [Attribute] format perform best.

Perplexity

Crawls schema.org JSON-LD directly from your product pages. If you don’t have JSON-LD, you’re invisible to Perplexity’s product recommendations. They specifically look for complete Product schema with Offer data.

Google AI Mode

Uses a combination of your Merchant Center feed, on-page JSON-LD, and crawled content. The richer your structured data, the more confidently Google’s AI can recommend your products.

The 5 most common JSON-LD mistakes

1. Missing identifiers

No GTIN, MPN, or SKU means AI agents can’t verify your product is real. This is the single biggest trust signal — without it, your product gets a low confidence score.

2. Vague titles

“Amazing Running Shoe” tells an AI nothing. “Acme Athletics Women’s Lightweight Running Shoe, Wide Width (D), Mesh Upper, 7.2oz” gives the AI everything it needs to match against specific queries.

3. Empty or generic descriptions

AI agents match descriptions to conversational buyer queries. “Great shoe for running” won’t match “best lightweight running shoe for women with wide feet under $150.” Attribute-dense descriptions do.

4. Missing availability data

Just saying “InStock” isn’t enough anymore. AI agents penalize merchants whose availability data is imprecise. Include exact quantities, handling times, and shipping estimates when possible.

5. No Q&A or conversational content

Google’s AI Commerce attributes include Q&A pairs and usage scenarios. Almost no merchants provide these, but they’re exactly what AI shopping agents use to match products to conversational queries.

How to check your JSON-LD

You can inspect any product page’s JSON-LD by:

  1. Right-clicking the page and selecting “View Page Source”
  2. Searching for application/ld+json
  3. Looking for a @type: "Product" block

Or use Lumio’s free AI Readiness Audit — we scan your entire catalog and score each product’s schema completeness across six dimensions, showing you exactly what’s missing and what to fix first.

The bottom line

JSON-LD isn’t optional anymore. It’s the language AI shopping agents speak. If your product pages don’t have complete, attribute-rich structured data, you’re invisible to the fastest-growing discovery channel in ecommerce.

The good news: fixing it is straightforward. Start with identifiers (GTINs), restructure your titles, and make sure every product page has complete Product/Offer JSON-LD. Or let Lumio do it for you.