Skip to main content
Guide 5 of 10 · 3 min read

How to Set Up Variants and Collections So AI Gets Them Right

When an AI agent reads your product data, it sees your variants as a list of option names, option values, SKUs, prices, and inventory levels. If those option names are "Option1" and "Option2," the agent has nothing to work with.

A customer asks "do you have this in blue?" — the AI agent checks your variant options. If it finds Color: Blue, it can answer. If it finds Option2: BL, it can't.

What AI agents see in your variants

Your /products.json feed exposes every product's variants and options. For each product, the agent gets:

  • Option names — the labels for each variant axis (e.g., "Size", "Color")
  • Option values — the specific choices (e.g., "Small", "Blue")
  • SKUs — unique identifiers for each variant
  • Prices — per-variant pricing
  • Inventory — stock availability per variant

The option names are the most important part. They tell the agent what dimension each variant represents.

Fix your option names

Shopify defaults to "Title" as the option name for products with no real variants. This is fine for single-variant products. But when you have actual variants, use clear, human-readable names:

Instead ofUse
Option1Size
Option2Color
Option3Material

And for option values:

Instead ofUse
BLBlue
S, M, LSmall, Medium, Large
01, 02, 03Cotton, Linen, Silk

AI agents read these option values as plain text. The clearer they are, the better the agent can match them to customer queries.

Add SKUs to every variant

SKUs (Stock Keeping Units) are unique identifiers for each specific variant of a product. Many stores leave these empty, but they matter for AI.

AI agents use SKUs to:

  • Precisely identify a specific variant in recommendations
  • Track inventory and availability accurately
  • Match products across different systems and platforms

A good SKU is descriptive and consistent. For example: TSHIRT-BLU-M for a blue medium t-shirt. But even if your SKU system is internal codes, having any SKU is better than having none.

How to check your SKUs

  1. Go to Products > Export in Shopify admin
  2. Export all products as CSV
  3. Open the CSV and check the Variant SKU column
  4. Look for empty cells — those are variants without SKUs

How to update variants in Shopify

  1. Go to Products in your Shopify admin
  2. Click on a product with variants
  3. Scroll to the Variants section
  4. Click on the option name (e.g., "Option1") to rename it
  5. Update each variant's SKU in the variant details
  6. Save

For bulk updates, the CSV export/import method works well. Update the Option1 Name, Option1 Value, and Variant SKU columns across all products.

Collections as category signals

Collections tell AI agents how your products relate to each other. A well-organized collection structure helps agents understand your product catalog.

Some tips:

  • Use descriptive collection names. "Men's Cotton T-Shirts" is more useful than "Summer Picks."
  • Create category-based collections. "T-Shirts", "Pants", "Accessories" — these help AI agents map products to standard categories.
  • Use automated collections where possible. Conditions based on product type, tags, or vendor keep collections accurate as you add products.
  • Don't rely only on manual collections. Manual collections can go stale as your catalog changes.

AI agents use collection titles and the products inside them to build a picture of what your store sells and how products are grouped. Clear collection names lead to better product recommendations.

Check your store

We're launching soon. Get notified when you can scan your store.