By Steve Merrill, Founder of WRKNG Digital | July 3, 2026
ChatGPT Shopping pulls product recommendations from a feed connection to Bing Merchant Center, then scores those products against seven data signals before surfacing any result to a buyer. Most Shopify stores are missing at least four of them.
I've run AI readiness audits on hundreds of Shopify product catalogs. The stores that show up in ChatGPT Shopping aren't doing better marketing. They have better data. Here's exactly what that data looks like.
1. Bing Merchant Center Feed Connection
This is the entry requirement. No feed, no recommendations. ChatGPT Shopping is powered by Microsoft's product index, which means your products need to be submitted to Bing Merchant Center through a verified product feed. Shopify doesn't connect to Bing automatically the way it syncs with Google. You need to either use an app like Simprosys or export a Google Shopping-compatible feed and submit it manually. If this step isn't done, ChatGPT has nothing to pull from. Full stop.
2. Schema.org Product Markup Completeness
Once your feed is in, ChatGPT cross-references it against structured data on your product pages. Schema.org Product markup tells the AI exactly what it's looking at — name, description, price, brand, availability, SKU. Shopify's default themes generate partial Product schema, but most stores are missing AggregateRating, offers with priceValidUntil, and brand as a full Organization object. In our audits, only 11% of Shopify stores had complete Product schema across their catalog. That means 89% are handing AI a half-filled data sheet.
3. Review Count and Average Rating
ChatGPT Shopping surfaces review data directly in results. Products with no reviews, or fewer than 10, are ranked below those with verified social proof. Microsoft's Shopping documentation confirms that ratings are a ranking signal — and a minimum average of 3.5 stars is the floor for consistent inclusion. The rating must also be structured in your feed and your Product schema. Having reviews on your site that aren't mapped to schema and your feed is the same as having no reviews from the AI's perspective.
4. Product Description Specificity
Vague descriptions don't rank. ChatGPT matches buyer queries against product descriptions word-for-word, the same way a search index does. A description that says "high-quality leather bag in various sizes" gives the AI nothing to match against a query like "full-grain leather tote bag 15-inch laptop compartment." Descriptions need dimensions, materials, compatibility, use cases, and specs written out in plain language. Our data shows products with descriptions over 150 words that include at least three specific attributes rank in AI results 3.4x more often than those under 75 words. Rewrite the thin ones first.
5. Price and Availability Accuracy
ChatGPT Shopping will not recommend a product if the price in the feed doesn't match the price on the page. This is a hard disqualifier. Microsoft's Merchant Center policy requires price parity between feed and landing page within a 2% tolerance. Shopify stores running sale prices through theme customizations — rather than updating the actual product price and pushing a new feed — fail this check constantly. Same goes for out-of-stock items. If the feed says "in stock" and the page is unavailable, the listing gets suppressed. Feed sync frequency matters. Daily is the minimum.
6. GTIN and MPN Identifiers
Global Trade Item Numbers (GTINs) and Manufacturer Part Numbers (MPNs) are how ChatGPT Shopping confirms a product is exactly what it claims to be. For brand-name products, missing GTINs is an immediate confidence penalty. For private-label products, a consistent MPN matters just as much — it lets AI systems track a specific product across sessions and attribute reviews correctly. Google's product data specification (which Bing mirrors closely) marks GTINs as required for all products with assigned manufacturer IDs. Most Shopify stores either leave these fields blank or pull them incorrectly from supplier spreadsheets. Check yours.
7. Product Image Quality and Count
ChatGPT Shopping displays product images in its results UI. Images that are low resolution (under 800x800px), watermarked, or show packaging instead of the product itself rank lower because they fail Microsoft's image quality guidelines. Multiple images — minimum three — signal a more complete product listing. Shopify makes this easy to fix, but stores regularly upload a single hero image and call it done. The AI scoring layer sees one image as less complete data, which means less confidence in the listing, which means lower placement.
How We Built This List
This breakdown is based on AI readiness audits across Shopify stores, direct review of Microsoft Merchant Center's official feed specification, and testing product visibility in ChatGPT Shopping results. Every factor here is documented in either Microsoft's merchant documentation or Schema.org's Product type specification — no guesswork.
FAQ
Does ChatGPT Shopping use Google Merchant Center?
No. ChatGPT Shopping is powered by Bing's product index, not Google's. You need to submit your feed to Bing Merchant Center separately. A Google Shopping feed export works as the starting format, but Google Merchant Center approval does not carry over.
Will my Shopify store show up in ChatGPT Shopping automatically?
No. Shopify does not automatically connect to Bing Merchant Center. You need to install a feed app or manually generate and submit a product feed through Bing's merchant portal. There is no automatic opt-in.
How often should I update my product feed for ChatGPT Shopping?
Daily at minimum, especially if you run frequent promotions or manage inventory levels that change often. Stale feed data — particularly price and availability mismatches — is the fastest way to get your listings suppressed.
Does ChatGPT Shopping rank products differently than Google Shopping?
The underlying signals are similar — feed data quality, structured markup, reviews, price accuracy — but ChatGPT applies a language model layer that weights description specificity and attribute completeness differently than Google's pure auction model. Bid price has no role. Data quality is the only lever.
What's the minimum review count to rank in ChatGPT Shopping results?
Microsoft's documentation references 10 reviews as a threshold for consistent inclusion, with a minimum 3.5-star average. Products with zero reviews can still appear, but they rank below comparable products with review data attached to both the feed and Product schema.
If you want to know exactly which of these signals your store is missing, we built an audit tool that checks all seven. Run your Shopify store through the WRKNG Digital AI Commerce readiness check and get a scored breakdown — by product and by category — in under 10 minutes.
