Resources
Practical, evidence-led guides for Shopify merchants whose sales depend on fit, sizing, compatibility and specs.
-
An AI assistant says your product is out of stock. It isn't. Here is why, and how to check
When ChatGPT, Gemini or another assistant tells a shopper your product is sold out while every variant is available, the cause is almost always which availability signal it read: your schema, your syndicated catalog, or one variant standing in for the whole product. Here is how to find which one, and fix it.
· 5 min read
-
How to check what ChatGPT, Gemini and Copilot say about your store (manually, and with a simulated check)
A single screenshot of one assistant answer tells you almost nothing. Here is the fresh-session method merchants use to check what ChatGPT, Gemini, Copilot and Perplexity say about their products, the buyer-intent questions worth asking, why the answers differ between sessions and apps, and what a repeatable simulated check adds on top.
· 9 min read
-
What AI shoppers get wrong about specialist products: 34 stores, 340 questions
We ran simulated AI shoppers against 34 public Shopify storefronts selling parts, tools, cameras and outdoor gear, using only the storefront tools every store already has. 45% of answers were fully right. Four of five classifiable misses were a single missing fact about a single product, not a failure to find it.
· 5 min read
-
Agentic Storefronts, explained for merchants: what Shopify sends to ChatGPT, Gemini, Copilot and Meta, and what it doesn't
Agentic Storefronts is on by default for eligible Shopify stores. It sends each product's title, description, options, images, price and availability to ChatGPT, Google AI Mode and Gemini, Microsoft Copilot and Meta. It does not send your metafields, size charts or compatibility tables. Here is what each channel does, what the catalog carries, and what that means when a shopper asks whether something fits.
· 12 min read
-
Which product data AI shopping agents can actually read from your store (and which they can't)
An AI shopping agent can reach your Shopify store through five surfaces: the product page, /products.json, the syndicated Shopify Catalog, the UCP catalog endpoint and the WebMCP storefront tools. Each returns a different slice of your product record. None returns a metafield. Here is what each one reads, with dates, and what to do about the facts that live in the fields nobody reads.
· 13 min read
-
Why AI quotes the wrong price for your variant (and how to make the right one readable)
A shopper asks an assistant what the case for their phone costs and hears the cheapest option's price, or a range. The product tool returned one variant and a price span, and the agent never asked for the option the shopper named. Here is the mechanism, a ten-minute check, and three fixes, cheapest first.
· 8 min read
-
The AI shopping tools every Liquid storefront got in August 2026: the ten tools, what they read, what they can't
Since August 2026 every Liquid storefront registers ten tools that a browser-based AI agent can call: search, browse, product details, cart, checkout, orders and a policy lookup. Here is what each one reads, which shopper questions that covers, which it doesn't, and how to see the list on your own store in one line of Chrome console.
· 11 min read
-
Should you turn off Agentic Storefronts? What opting out does, and what fixing does instead
You can stop ChatGPT, Copilot, Google and Meta from reading your products through Shopify Catalog, per channel, under Sales channels → Agentic. You cannot opt out of the catalog itself, the change takes up to seven days, and assistants can still find your products by crawling. Here is exactly what the switch changes, what it leaves in place, when turning it off is the right call, and what to do if the real worry is wrong answers.
· 10 min read
-
Size charts AI shoppers can read: data, not images
A shopper tells an assistant their head measures 56 cm and asks which helmet size to order. If your size chart is an image, a PDF, a pop-up page or a metafield no tool returns, the assistant cannot answer from it. Here is what each chart format looks like to an agent, the data model that works, how to check your own store, and what a read-only sizing tool adds.
· 11 min read
-
"Will it fit?" Compatibility data that AI assistants can use
A shopper asks an assistant whether these pads fit their calipers, whether this rotor fits their truck, whether this protector fits the 15.4-inch model. The merchant already knows. The answer sits in a metafield, a spec table or a fitment app's own database, and none of the storefront surfaces an agent reads returns it. Here is the question shape, what our checks found, the data model that makes fit answerable, and where fitment apps still belong.
· 11 min read
-
Shopify's readiness scanner, free AI checkers and StoreKnows: what each one actually tests
Five tools will tell you whether your store is ready for AI shoppers, and they measure five different things. Shopify's scanner and Craftshift's checker test files, schema and endpoints. AI Catalog Score grades catalog fields. Verity Score logs what four models say. StoreKnows grades answers to fit, size, spec and compatibility questions against your own product data. Here is what each one checks, what each costs, and when the free one is enough.
· 12 min read
-
Same store, same questions, better answers: what changed when product data became readable
We asked the same 14 shopper questions of the same 144-product development store twice: once with only the storefront's native tools, once with five read-only tools serving the product's metafields beside them. Google's Gemini 3.8 Flash went from 7 to 13 fully right answers; OpenAI's GPT-5.4 mini from 4 to 8. One store, one run each, and the limits that go with that.
· 10 min read