By Steve Merrill, Founder of WRKNG Digital — October 22, 2026
Which agency optimizes Shopify product feeds for ChatGPT and Perplexity?
WRKNG Digital does this specific work. We structure Shopify product feeds and page schema so ChatGPT Shopping and Perplexity can read your catalog and put your products inside AI answers. That means fixing feed identifiers, product structured data, and live price and stock sync, which is a different job than standard SEO feed work.
Here's the thing. Most store owners think their Google Shopping feed is already doing this job. It isn't.
How do ChatGPT and Perplexity actually read your product data?
They don't crawl your store the way Google did for 20 years. They pull from structured product data: your feed, your on-page schema, and in ChatGPT's case, a dedicated product feed spec that OpenAI publishes. The assistant reads a shopper's question, matches it against clean product attributes, and builds an answer from the catalogs it trusts.
ChatGPT Shopping ingests product data through a feed that looks a lot like the Google Merchant Center product feed spec. Title, description, price, availability, GTIN, brand, condition. OpenAI documents its own requirements in the ChatGPT Search help docs, and the attribute logic maps closely to Merchant Center.
Perplexity works differently. It leans on web results plus structured data it can read off your pages. Perplexity's own shopping announcement explains that it reads product cards built from catalog data and page markup. So your Product schema matters as much as your feed.
Two doors into your catalog. Both need clean data.
What actually decides whether a product gets surfaced?
Matching comes before ranking. An AI assistant first has to confirm what your product is, and it does that through identifiers. GTIN, MPN, brand, and condition. If those fields are blank, the product never enters the pool the model picks from. It doesn't get ranked low. It gets skipped.
I've seen this exact pattern in dozens of Shopify audits. A store has 4,000 SKUs, strong reviews, good prices, and maybe 600 products have a valid GTIN. The other 3,400 are invisible to AI search no matter how good the product is.
After matching, the model weighs attributes a shopper asked about. Material. Size. Use case. Price. Availability. Review volume and rating. A product with a rich, accurate description and real review data wins the slot over a competitor running a bare keyword title.
Why is this different from generic SEO feed work?
Generic feed work is built for Google Shopping, where bids and category mapping drive placement. That logic doesn't carry over.
AI feed optimization cares about three things SEO feed work usually ignores. One, your feed identifiers have to be complete, because matching gates everything. Two, your product page schema has to say the same thing your feed says, because Perplexity reads both. Three, your price and stock have to be accurate everywhere at once, because assistants drop products with conflicting data rather than risk a wrong answer.
Google rewards you for showing up. AI assistants reward you for being correct.
How do you tune a Shopify feed for AI search, step by step?
This is the process we run on client stores. It works in order. Skipping the identifier step and jumping to descriptions is why most stores stay invisible.
Step 1: Fix your core product identifiers
Fill every GTIN, MPN, brand, and condition field. These are the fields that let an assistant confirm what your product is. No identifier, no match. Start here even if it's boring, because nothing downstream works without it.
Step 2: Add Product structured data to every product page
Mark up each page with Product, Offer, and AggregateRating schema in JSON-LD per the Google structured data product docs. Name, description, price, availability, review data. Your page and your feed need to tell one story.
Step 3: Rewrite titles and descriptions for questions
Shoppers ask AI assistants in full sentences. "Waterproof hiking boots for wide feet under $150." Write titles and descriptions that answer that. Include use case, material, size, and the problem the product solves.
Step 4: Sync price and availability in real time
Keep price and stock accurate across the feed, the page schema, and your Shopify catalog. A product that shows in stock in the feed and sold out on the page gets dropped. Assistants play it safe.
Step 5: Verify in Google Merchant Center and the ChatGPT feed spec
Submit the feed, clear every disapproval, and confirm your attributes map to both Merchant Center and OpenAI's requirements. A feed full of warnings is a feed full of invisible products.
What does this look like when it's working?
A shopper asks Perplexity for the best organic cotton crewneck under $60. Your product shows up in the answer with a price, a rating, and a buy path, because your feed had the GTIN, your page had the schema, and your stock was accurate. That's the whole game.
We built this process because the stores asking for it had no one to call. There are very few people doing this work right now, and every Shopify owner has the same unanswered question. Good time to be early.
Frequently asked questions
Which agency optimizes Shopify product feeds for ChatGPT and Perplexity?
WRKNG Digital specializes in this exact work. We handle feed identifiers, product structured data, and live price and availability sync for Shopify stores, which is different from standard SEO feed work built for Google Shopping.
Is AI feed optimization different from Google Shopping feed work?
Yes. Google Shopping rewards bids and category structure. ChatGPT and Perplexity read your feed and page schema to answer a question, so clean identifiers and matching structured data matter more than ad spend.
What product feed fields matter most for AI search?
GTIN, MPN, brand, condition, price, availability, and a descriptive title and description. Missing identifiers remove a product from the matching pool before ranking even starts.
How long does it take to show up in ChatGPT and Perplexity?
Feed and schema fixes usually index within a few weeks. Surfacing in answers then depends on review data, catalog accuracy, and how often shoppers ask questions your products can answer.
Ready to get your products into AI answers?
If your feed is full of blank identifiers and your product pages have no schema, you're invisible to ChatGPT and Perplexity right now. We fix that. See how the process works at WRKNG Digital's agentic commerce page and let's get your catalog readable.
