Is Your Product Data AI-Ready? The Complete Checklist

AI shopping agents do not browse your store the way humans do. They do not squint at product photos, skim marketing copy, or click through size guides. They parse structured data fields — and if a field is missing, malformed, or inconsistent, your product is invisible to them. Period.

This checklist covers every data point your products need to be discoverable by AI agents like ChatGPT, Claude, Perplexity, and Google's AI Overviews. Use it as a literal punch list: go through each item, check it off, and fix what is broken. Your GEO score will improve with every fix.

JSON-LD Structured Data Checklist

JSON-LD is the preferred format for structured product data. Every product page on your store should include a single <script type="application/ld+json"> block containing a complete Product schema. Here is what "complete" means.

Required Fields (Missing These = Invisible)

• [ ] `@context`: Set to "https://schema.org"
• [ ] `@type`: Set to "Product"
• [ ] `name`: Full product name including brand and key variant (e.g., "Nike Air Max 270 - Men's Black/White"). Do not stuff keywords. Keep it under 150 characters.
• [ ] `description`: Minimum 150 words. Include materials, dimensions, weight, compatibility, and primary use cases. No HTML tags inside the JSON-LD value — plain text only. This is the single most commonly deficient field.
• [ ] `sku`: Your internal SKU identifier. Must be unique across your catalog.
• [ ] `image`: At least one URL. Ideally an array of 3+ image URLs. Images should be at least 800x800 pixels, on a white or neutral background, and show the product clearly from multiple angles.
• [ ] `brand.name`: The manufacturer or brand name, nested inside a Brand object.
• [ ] `offers.price`: Current price as a decimal string (e.g., "79.99"). Must reflect the actual price the customer pays, not a "compare at" price.
• [ ] `offers.priceCurrency`: ISO 4217 currency code (e.g., "USD", "EUR", "GBP").
• [ ] `offers.availability`: Schema.org availability enum. Must be one of: https://schema.org/InStock, https://schema.org/OutOfStock, https://schema.org/PreOrder, https://schema.org/BackOrder. Must reflect real-time inventory status — not hardcoded.
• [ ] `offers.url`: Canonical URL for this product page.
• [ ] `offers.itemCondition`: One of https://schema.org/NewCondition, https://schema.org/UsedCondition, https://schema.org/RefurbishedCondition.

High-Impact Optional Fields (Missing These = Disadvantaged)

• [ ] `gtin13` or `gtin`: The product's Global Trade Item Number (GTIN-13 or GTIN-14). This is the single highest-impact field for AI shopping visibility. Products with valid GTINs appear up to 60% more frequently in AI agent recommendations because agents can cross-reference them against global product databases. If you sell branded products, get the GTIN from your supplier. If you manufacture your own products, register with GS1 ($250/year for small businesses).
• [ ] `mpn`: Manufacturer Part Number. If the product has no GTIN, the combination of mpn + brand.name serves as an alternative identifier.
• [ ] `aggregateRating`: Include ratingValue (e.g., 4.7), reviewCount (e.g., 342), and bestRating (e.g., 5). AI agents heavily weight products with rating data — it serves as a trust signal and is displayed prominently in recommendations.
• [ ] `review`: At least 1-3 structured Review objects with author, reviewRating, and reviewBody. Agents use review text to extract product attributes that may not be in your description.
• [ ] `offers.shippingDetails`: Include shippingRate (cost), handlingTime, and transitTime. AI agents increasingly factor shipping speed and cost into recommendations.
• [ ] `offers.hasMerchantReturnPolicy`: Link to a MerchantReturnPolicy schema on your returns page. Transparent return policies boost agent trust scores.
• [ ] `color`: Product color. Used for attribute-based filtering.
• [ ] `size`: Product size. Used for filtering — missing this means size-specific queries skip your product entirely.
• [ ] `material`: What the product is made of. Increasingly used in eco-conscious and technical queries.
• [ ] `weight`: Product weight with unit. Critical for categories like outdoor gear, shipping calculations.
• [ ] `additionalProperty`: Use this for custom attributes that do not have dedicated schema.org fields (e.g., "battery life", "water resistance rating", "thread count").

Product Feed Checklist

Your product feed (Google Merchant Center, Bing Merchant Center, and/or Facebook Catalog) is a bulk data channel that AI systems use alongside on-page structured data. Here is what a complete feed entry needs.

Required Feed Attributes

• [ ] `id`: Unique product identifier. Use your SKU. Must be stable — do not change IDs when updating products.
• [ ] `title`: Match your on-page product name. Include brand + product type + key differentiator. Max 150 characters.
• [ ] `description`: Match your on-page description. Discrepancies between feed and page trigger trust penalties. Min 150 words.
• [ ] `link`: Canonical product page URL. Must be HTTPS. Must resolve to a live page.
• [ ] `image_link`: Primary product image URL. Min 800x800 pixels. No watermarks, no promotional overlays.
• [ ] `price`: Current price with currency (e.g., "79.99 USD"). Must match on-page price.
• [ ] `availability`: One of in_stock, out_of_stock, preorder, backorder. Must match on-page status.
• [ ] `brand`: Brand name. Required for all branded products.
• [ ] `condition`: One of new, refurbished, used.
• [ ] `gtin`: 8, 12, 13, or 14-digit GTIN. Required for all products with manufacturer-assigned GTINs. Google will disapprove products missing GTINs for certain categories.

Recommended Feed Attributes

• [ ] `additional_image_link`: Up to 10 additional image URLs. More images = better visibility.
• [ ] `sale_price`: If the product is on sale, include the sale price and the sale_price_effective_date.
• [ ] `google_product_category`: Google's taxonomy ID for the product category. Be as specific as possible (e.g., "Apparel & Accessories > Clothing > Outerwear > Coats & Jackets" rather than just "Apparel").
• [ ] `product_type`: Your own category taxonomy. Helps with search relevance.
• [ ] `color`: Product color. Use standard color names, not creative names ("Midnight Dream" should be "Navy Blue").
• [ ] `size`: Use standard sizing (S, M, L, XL or numeric).
• [ ] `material`: Primary material.
• [ ] `shipping`: Shipping cost and speed by region. Agents use this for delivery estimates.
• [ ] `tax`: Tax information by region.
• [ ] `mpn`: Manufacturer Part Number. Required if no GTIN is available.
• [ ] `multipack`: If selling a multipack, indicate the count.
• [ ] `is_bundle`: Set to true if the product is a bundle.
• [ ] `age_group` / `gender`: Required for apparel and accessories.
• [ ] `item_group_id`: Group variants (sizes, colors) under a single parent. Critical for variant-rich catalogs.

Feed Quality Checks

• [ ] Feed refreshes at least every 4 hours. Hourly is better. Real-time is best.
• [ ] Zero disapproved items in Google Merchant Center. Check weekly.
• [ ] No price/availability mismatches between feed and live site. This is the #1 cause of feed errors.
• [ ] No duplicate IDs or duplicate GTINs in the feed.
• [ ] All URLs resolve to live pages (no 404s, no redirects to homepage).

Product Description Guidelines

AI agents parse product descriptions to extract attributes, understand use cases, and generate recommendation text. Thin or vague descriptions are a liability.

Length and Structure

• [ ] Minimum 150 words per product description. 200-300 words is optimal for most categories.
• [ ] Lead with the key selling proposition in the first sentence. Agents may truncate descriptions.
• [ ] Use natural language paragraphs, not keyword-stuffed bullet lists. Agents use NLP — they understand well-written prose better than fragmented bullet points.
• [ ] Include a specification section with quantified attributes: dimensions, weight, capacity, power, etc.

Content Requirements

• [ ] Materials and construction: What is it made of? How is it made? (e.g., "3-layer GORE-TEX fabric with sealed seams and YKK zippers")
• [ ] Dimensions and weight: Exact measurements with units. "Lightweight" is meaningless; "weighs 14 oz / 397g" is useful.
• [ ] Compatibility: What does it work with? What does it fit? (e.g., "Compatible with iPhone 14/15/16 series, fits cases up to 3mm thick")
• [ ] Use cases: When and where to use it. (e.g., "Designed for 3-season backpacking in wet climates")
• [ ] Care instructions: If applicable. Agents surface this for apparel and textile queries.
• [ ] What is in the box: List all included items. Reduces ambiguity.

What to Avoid

• [ ] No HTML tags in the JSON-LD description value. Plain text only.
• [ ] No promotional language that expires: "Sale ends Friday!" becomes stale data.
• [ ] No competitor references: "Better than Brand X" is not parseable by agents.
• [ ] No ALL CAPS or excessive punctuation!!! Agents may flag this as low-quality content.
• [ ] No duplicate descriptions across products. Each product needs unique content.

Image Requirements

AI agents display images in their recommendations, and they increasingly use image understanding to validate product data.

Technical Specs

• [ ] Minimum resolution: 800x800 pixels. 1200x1200 or higher is preferred for zoom capability.
• [ ] File format: JPEG or PNG. WebP is acceptable but ensure a JPEG fallback for maximum compatibility.
• [ ] File size: Under 5MB per image. Optimize with lossy compression at quality 80-85%.
• [ ] Aspect ratio: Square (1:1) for the primary image. Additional images can vary.
• [ ] Background: White or light neutral background for the primary image. Lifestyle shots are valuable as secondary images.

Content Requirements

• [ ] Product clearly visible: No obstructions, no excessive whitespace, no tiny product in a large frame.
• [ ] No watermarks, logos, or promotional overlays on the primary image.
• [ ] Multiple angles: Front, back, side, detail close-up. Each as a separate image URL in both JSON-LD and feed.
• [ ] Variant images: If you sell color/pattern variants, each variant should have its own image.

Technical Accessibility Checklist

Your data can be perfect but unreachable. This section covers the technical requirements for AI agents to access it.

Crawl Access

• [ ] robots.txt allows AI crawlers: Verify that ChatGPT-User, Googlebot, Bingbot, PerplexityBot, ClaudeBot, and anthropic-ai are not blocked on product pages.
• [ ] No noindex tags on product pages: Check both meta robots tags and X-Robots-Tag HTTP headers.
• [ ] XML sitemap is current: Includes all active product URLs, uses accurate <lastmod> dates, and is submitted to Google Search Console and Bing Webmaster Tools.

Page Performance

• [ ] JSON-LD renders without JavaScript execution: Some SPAs require JS to render JSON-LD, which many crawlers skip. Verify that your structured data is in the initial HTML response.
• [ ] Page loads in under 3 seconds: AI crawlers have timeout thresholds. Server-side rendering or static generation is preferred over client-side rendering.
• [ ] HTTPS: Non-negotiable. HTTP-only stores receive zero technical accessibility score.

MCP Server (Bonus)

• [ ] MCP server deployed and accessible: Real-time product queries via Model Context Protocol.
• [ ] `/.well-known/mcp.json` manifest published: Tells agents where your MCP server is and what it offers.
• [ ] Rate limiting configured: Prevent abuse without blocking legitimate agent traffic.

Data Consistency Checks

The worst thing you can do is have conflicting data across channels. AI agents cross-reference sources and penalize inconsistencies.

• [ ] Price matches across JSON-LD, product feed, and live page display.
• [ ] Availability matches across JSON-LD, product feed, and actual inventory.
• [ ] Product titles match across JSON-LD and product feed (minor formatting differences are acceptable).
• [ ] Images match across JSON-LD and product feed (same URLs or equivalent content).
• [ ] GTIN is identical across JSON-LD, product feed, and any other data source.

Running Your Audit

The fastest way to audit your product data against this checklist is to run a SignalixIQ scan. The scan:

1. Crawls your product pages and extracts JSON-LD
2. Validates every field against schema.org requirements
3. Checks your product feeds for completeness and consistency
4. Tests crawl accessibility and page performance
5. Generates a GEO score with specific fix recommendations

For manual auditing, use these tools:

• Google Rich Results Test: Validates your JSON-LD markup
• Google Merchant Center diagnostics: Shows feed errors and disapprovals
• Schema.org validator: Checks schema compliance
• PageSpeed Insights: Measures page load performance

Work through this checklist systematically, starting with the required fields. Every field you add, every inconsistency you fix, and every image you optimize makes your products more discoverable by the AI agents that are rapidly becoming a primary shopping channel. Read our ChatGPT Shopping guide for platform-specific integration steps.

Frequently Asked Questions

Related Articles