ChatGPT ads are open to every Shopify merchant (what to know)

Your Shopify product title is the ad headline and your product description is the body copy. Your taxonomy category and the metafields that come with it determine which queries your product is eligible to match against.

Last month, OpenAI confirmed this by launching product feed ads inside ChatGPT. Retailers connect their existing structured product catalog to the platform, apply eligibility filters, and the system automatically generates sponsored placements based on the product names, images, and catalog attributes it finds there. No separate ad creative or copy to write. The placement is built from the feed.

The feed format ChatGPT accepts is the same format already sent to Google Merchant Center. One feed - two surfaces.

The self-serve Ads Manager that makes this accessible to any brand: ads.openai.com. Advertisers can register, set a budget, launch campaigns, and view performance directly. CPC bidding is now supported alongside the earlier CPM model.

How the matching works

The placement appears below ChatGPT's response as a clearly labelled sponsored card. Shopping carousels surface when users are in purchase-intent conversations - describing what they need, comparing options, asking about a category.

There is no keyword to bid on. The match happens between the natural language in the conversation and the structured data in the product feed.

A user asking "cordless drill with brushless motor, under $200, for home renovation" is not entering a search term - they are describing a specific requirement, and the platform matches it against the structured attributes in your catalog.

If the brushless motor specification is only mentioned in your product description, the system may not surface it reliably as a filterable attribute. If it is set as a variant option or a metafield, the match is direct.

The quality of the match depends entirely on how structured the product data is.

The eligibility gate that most brands will fail

Before OpenAI accepts a brand's full catalog for paid placement, they require a 100 product sample feed - for review. If the sample passes their quality standards, the catalog is connected. If it fails, the brand cannot run product feed ads until the data is fixed.

Google Merchant Center and Meta do not gate the catalog this way. Both review products after submission and disapprove individual items. Neither blocks a brand from bidding because the first 100 SKUs have thin or missing data.

OpenAI is filtering at the source → your Shopify store.

The sponsored card surfaces directly beneath a conversational answer, where a poorly matched product recommendation reflects on ChatGPT's answer quality before it reflects on the advertiser's creative. The quality control has to happen upstream of the auction.

What the sample review checks: complete product titles that describe the product specifically rather than leading with brand names, accurate and specific descriptions, correct Standard Product Taxonomy category assigned at leaf level, structured metafields for the product type (material, color, dimensions, use case specifics), per-variant availability tracked at the variant level, and images at minimum - 256x256px.

Your Shopify catalog is already connected

Since March, every eligible Shopify merchant's products have been discoverable inside ChatGPT conversations through Agentic Storefronts. No installation required - Shopify Catalog syndicates the data automatically through the ACP (Agentic Commerce Protocol). The same catalog that determines whether your products show up in organic ChatGPT responses is the feed you will submit for paid placement.

If you run a query for your category in ChatGPT right now and your products do not appear despite being the right match, the catalog data is the reason. And the 100-product sample review will surface the same issues for the same reasons.

The organic test tells you exactly what the paid eligibility review will find.

I built a free tool that shows which fields are missing across your catalog. If you want to check before connecting to Ads Manager → Take a free AI audit

Three things to fix before

Category assignment first - products without a leaf-level Standard Product Taxonomy category have no metafields populated, no taxonomy node for the system to classify them under, and no eligibility for category-level matching. After thousands of store audits, this is the most common gap we see - not one or two products, but the majority of the catalog. The fix is in Shopify admin - Products → each product → Category field.

Structured attributes second - once the correct category is assigned, the metafields for that product type appear automatically. Fill them. Material, color, dimensions, fragrance-free status, brushless motor specification - whatever applies to the product. These are the fields the system filters by when matching a natural language query to a specific product. Attributes mentioned only in the PDP copy/description are not structured data and will not match against attribute-based queries.

Description specificity third - "Premium quality crafted for the modern home" gives the system nothing to match against a specific query. "100% linen shirt, relaxed fit, machine washable, available in natural and stone" gives it material, cut, care, and colorway - enough to match queries that mention any of those attributes.

Enrichment Process at Scale with AI

Catalog Genius reads your catalog to find agent discovery gaps, then assigns the correct Product Taxonomy category to each product, and enriches the metafields.

Install Catalog Genius on Shopify - first 20 products free

- Ankit