How to Prepare Your Catalog for Shopify Catalog and AI Channels

Published 5 min read
How to Prepare Your Catalog for Shopify Catalog and AI Channels

When a shopper asks an AI assistant for “a navy winter jacket under $200,” the answer comes from product data, not from your homepage design. Shopify now feeds that data to AI shopping channels through Shopify Catalog and Agentic Storefronts. This guide shows how to make your products eligible, and how to improve the chance they’re actually recommended. Features, channels, and availability are rolling out in phases, often US-first, so check your admin for what applies to your store.

How it works

Shopify Catalog combines the eligible products from Shopify stores into one structured catalog: title, description, images, price, inventory, shipping, variants, and policies. Agentic Storefronts then make those products available in AI channels such as ChatGPT and Microsoft Copilot, with the list of channels changing as the rollout continues. Merchants reportedly stay the merchant of record, and in some channels checkout happens on your own store.

1. ELIGIBLE
Product meets the gates
2. RETURNED
Matches the shopper’s query
3. RANKED
Data quality, relevance, availability, price, engagement

Being present isn’t the same as being chosen. Data quality, relevance, availability, price, and engagement are all reported as factors. Your job is to be clean, complete, and accurate.

1

Check the eligibility gates

Products need:

✓
A title and at least one image
✓
A price above zero (free products are excluded)
✓
Publication to the Online Store (or Hydrogen or Headless channel) with a valid product URL
✓
To not be unlisted or hidden from search engines
✓
An allowed category. Uncategorized products and some restricted categories are reportedly excluded.

The store itself needs to be on a paid plan (Starter or higher) and have no storefront password. It also generally needs to be B2B-eligible and in good standing, and some channels require selling to customers in specific countries, such as the US. Confirm the current rules in Shopify’s Help Center.

2

Categorize every product

Assign a category from Shopify’s Standard Product Taxonomy to every product. Category helps AI systems decide whether a product is a candidate for a given question, and it connects to filters and to Google’s categories. Make sure the product information agrees with the category you pick.

3

Add identifiers

•
Brand: correct vendor name
•
GTIN, UPC, or EAN barcodes wherever your products have them
•
MPN for private-label or custom items
•
Accurate SKUs

Products with missing identifiers, brand, image, or description are at higher risk of being skipped in recommendations.

4

Write titles for machines and people

Titles should say what the product is, in plain words:

Weak
“Winter Vibes Jacket”
Better
“Men’s Navy Insulated Winter Jacket, Waterproof”

Include brand, product type, and the main distinguishing attributes. Don’t stuff keywords or add promotional claims.

5

Put the facts in the description

Agents read attributes from fields and from the description. Include:

✓
Material, dimensions, weight, capacity
✓
Size and fit information
✓
Compatibility, care, and ingredients where relevant
✓
Intended use and who it’s for
✓
Warranty and what’s in the box

Be specific and consistent. Avoid unsupported claims, especially for health, safety, or performance.

6

Fix images

✓
Sharp main image on a clean background
✓
Several angles, plus scale or lifestyle images
✓
Images that match each variant
✓
Descriptive alt text
7

Structure variants properly

Keep variants of one product together under a single product, with clear option names such as “Navy / Medium.” Splitting one product into many separate listings fragments how agents see it. Remember that Shopify allows three options and up to 2,048 variants per product.

8

Keep prices and compare-at prices honest

Shopify reportedly began sharing compare-at prices with AI shopping channels in October 2026, returning them as a list price. If your “was” prices are stale or inflated, agents may repeat them. Review them before Black Friday.

9

Keep inventory accurate

Availability is a ranking factor, and a recommended product that turns out to be out of stock hurts trust. Sync inventory, mark sold-out items correctly, and avoid selling stock you don’t have.

10

Complete your policies

Shipping, return and refund, terms, and privacy policies should be complete and consistent. They’re both a requirement for some channels and a trust signal. Shopify can generate draft policies, but review them so they match what you actually do.

11

Add structured data and allow crawlers

✓
Make sure product pages have accurate Product structured data (name, image, price, currency, availability, brand, identifiers)
✓
Avoid duplicate or conflicting schema from your theme and apps
✓
Check that your robots settings aren’t unintentionally blocking the crawlers you want to allow
✓
If you use Google Merchant Center, keep your feed healthy and consistent with your pages
12

Use catalog mapping where it helps

Shopify lets you choose which data source feeds the title, description, and category that agents see, such as product attributes, metafields, or metaobject references. That lets you keep a clean agent-facing description without changing your storefront copy. Check availability for your plan.

13

Decide which channels you want

In your admin’s sales channel settings, you can manage which AI channels receive your products and review checkout preferences. Opting out of a channel is possible, though that may not remove all visibility, since AI systems can also find public pages. Think about margins, return handling, and customer service before enabling checkout inside agents. You’ll need to be the store owner or have the right staff permissions to manage these settings.

14

Monitor and improve

•
Review orders attributed to AI channels
•
Search for your products in AI assistants as a customer would, and check what they say
•
Fix wrong or missing facts at the source
•
Re-audit after major catalog changes

What to fix first

1
Eligibility gates: published, priced, not hidden, categorized
2
Barcodes, brand, and clean titles on best sellers
3
Complete descriptions with real attributes
4
Honest prices and accurate inventory
5
Policies and structured data
6
Everything else, product by product
Start with your top 20 products, not the whole catalog.

Quick checklist

✓
Every product published, priced above zero, and visible to search engines
✓
Taxonomy category assigned to every product
✓
Brand and GTIN or MPN on every product
✓
Clear, factual titles and descriptions
✓
Good images with alt text
✓
Variants grouped correctly
✓
Compare-at prices honest and current
✓
Inventory accurate
✓
Policies complete
✓
Structured data checked, with no duplicates
✓
Agentic channel settings reviewed
✓
AI assistants tested with real queries

Final thoughts

In AI shopping, your product data is your storefront. The stores that get recommended tend to be the ones with complete, consistent, accurate information. Clean up your top products first, keep your facts honest, and check regularly how assistants describe what you sell.