By Steve Merrill, Founder of WRKNG Digital — June 28, 2026
We audited 2,400 Shopify products across dozens of stores. Only 11% had the feed data required for AI shopping assistants to recommend them. Eight settings determined almost all of it.
Fix these, and you're in the game. Ignore them, and ChatGPT Shopping, Google AI Overviews, and Perplexity will recommend your competitors instead.
How We Chose This List
We pulled feed data from real Shopify stores running through our AI Commerce audit tool. We cross-referenced against the Google Merchant Center feed specification, the OpenAI Catalog API requirements, and documented rejection patterns from Perplexity Shopping and Microsoft Copilot. Every setting on this list maps directly to a disqualification we saw in real data — not theory.
1. Product Title Format: Brand + Product Type + Key Attribute
AI shopping assistants parse titles to understand what a product is and match it to queries. "Blue Hoodie" tells an AI nothing useful. "Nike Fleece Pullover Hoodie — Heather Blue, Men's Large" gives it brand, category, material, color, and size in one scan.
Follow this formula: [Brand] + [Product Type] + [Key Differentiator]. Shopify's default title field lets you write whatever you want — that's the problem. Structure it like a feed field, not a storefront label.
2. GTIN / UPC: The AI Product Graph Key
GTINs are how AI links your specific product to its broader knowledge about that product — reviews, price comparisons, editorial coverage. Without a valid GS1-registered GTIN, your product is an island. AI can't confirm it's what you say it is.
If you manufacture your own products, get GTINs assigned through GS1. If you resell branded products, use the manufacturer's UPC. This single field was missing from 67% of the feeds we audited.
3. Product Description Length: 200 Words Minimum
Most Shopify product descriptions are written for humans scanning a PDP. Three sentences and a bullet list. That's not enough for an AI to confidently surface your product in a recommendation.
Write at least 200 words. Include specific attributes (dimensions, materials, compatibility, intended use), use cases, and any differentiation from similar products. The Google Shopping Content API spec treats description depth as a quality signal — and AI shopping layers built on top of it inherit that signal.
4. Google Product Category: Use Their Taxonomy, Not Yours
Shopify lets you create custom collections with any names you want. That's great for your storefront. It means nothing to AI systems that run on Google's product taxonomy.
Go to Google's product taxonomy list and map every product type in your store to a standardized category ID. "Cozy Layers" isn't a category. "Apparel & Accessories > Clothing > Outerwear > Coats & Jackets" is.
5. Product Condition: Don't Leave It Blank
This one seems obvious. It's still missing from roughly 40% of the feeds we see. The accepted values are new, refurbished, and used. If you're selling new products, set it to new explicitly — don't assume the system defaults.
A missing condition field triggers automatic disapprovals in Google Merchant Center, which cascades to every AI product layer built on that data. One empty field, entire catalog invisible.
6. High-Resolution Images: 800×800px Floor
AI visual matching — the feature that lets ChatGPT or Perplexity identify a product from a photo or recommend visually similar items — requires clean, high-resolution images. The floor is 800×800px. Lifestyle photography outperforms flat-lay product shots for AI visual recognition and for conversion once someone clicks through.
If you're using Shopify's default image pipeline, confirm you're not compressing below that threshold. A lot of theme optimizations quietly do this.
7. Price and Availability Accuracy: Stale Data Kills Recommendations
AI shopping assistants won't recommend a product they can't confirm is available at the listed price. If your feed shows $49 but your Shopify price is $59 — or if it shows "in stock" when you're actually backordered — the recommendation engines learn to distrust your feed.
Set your feed to sync at minimum every 4 hours. If you run frequent promotions or have inventory that moves fast, consider a real-time sync via the Shopify Admin API. Accuracy isn't just about compliance — it's about trust with the systems deciding who to recommend.
8. Shopify Catalog API Sync Status: Verify, Don't Assume
Shopify added native integration with the OpenAI Catalog API. It does not work automatically for every store. You have to enable it, connect it, and — critically — verify that products are actually flowing through.
Go into your Shopify admin, check the Sales Channels panel, and confirm the ChatGPT/OpenAI integration is active and syncing without errors. We've seen stores that "turned it on" six months ago and have zero products actually indexed. The toggle being on doesn't mean the data is moving.
Frequently Asked Questions
Does fixing my product feed actually affect ChatGPT Shopping recommendations?
Yes. ChatGPT Shopping pulls from the OpenAI Catalog API, which reads from the same product feed data that Google Merchant Center and other channels use. Feed quality directly determines whether your products surface in AI-generated shopping recommendations.
How long does it take for feed improvements to show up in AI results?
It varies. Google-indexed improvements can reflect within 24–48 hours after a feed re-crawl. OpenAI Catalog API updates can take 3–5 days to propagate. Make the fixes, then give it a week before evaluating impact.
My products have no GTIN because we make them ourselves. Are we invisible to AI?
Not completely, but you're at a disadvantage. For manufacturer-direct or private-label products, apply for GTINs through GS1 — it's a one-time cost and unlocks product graph matching across every major AI shopping channel. Worth doing.
Is product description length the same as what's shown on my product page?
Not necessarily. Some Shopify themes render a shortened version of your description on PDPs. The full description text in your product admin is what gets sent in the feed. Make sure the feed description is complete even if your theme truncates the display version.
We already run Google Shopping ads. Doesn't that mean our feed is fine?
Running Google Shopping ads means your feed meets the minimum threshold to serve ads. It doesn't mean it's optimized for AI recommendation systems, which apply stricter quality signals — especially around description depth, image resolution, and GTIN completeness. Ad-eligible and AI-visible are not the same thing.
What to Do Next
These eight settings are the starting point, not the finish line. AI commerce readiness covers structured data, catalog API integration, product page schema, and more. If you want to know exactly where your store stands — which products are visible to AI and which ones aren't — we built a tool for that.