Best Practices for OpenAI Commerce Extension

✅ Always provide complete seller information—AI platforms use this to answer customer questions

✅ Use "Only Linked Simple Products" for configurable products so AI can recommend specific variants

✅ Map all available product identifiers (GTIN, Brand, MPN) for better product matching

✅ Include detailed product descriptions that answer common customer questions

✅ Set up automatic feed generation via cron to keep data current

✅ Create dedicated CMS pages for policies (privacy, terms, returns)

❌ Don't export configurable parent products—customers can't purchase them directly

❌ Don't leave seller data empty—it's required for proper platform integration

❌ Don't use HTML in descriptions—it gets stripped and may cause formatting issues

❌ Don't include out-of-stock products unless you have a strategy for them

❌ Don't forget to set your return window—AI may reference this when answering questions

Use case: Store selling simple products only (no variants).

Configuration:

Seller Data:

• Seller Name: Your Business Name
• Privacy Policy: Select your privacy page
• Terms of Service: Select your terms page
• Return Policy: Select your returns page
• Return Window: 30 (or your actual policy)

Attributes:

• ID: SKU
• Title: Product Name
• Description: Description
• GTIN: Your EAN attribute
• Brand: Manufacturer

Filter Options:

• Visibility: Catalog, Catalog/Search
• Exclude Out of Stock: Yes

Use case: Apparel store with size/color variants.

Configuration:

Attributes:

• All standard attributes mapped
• Color: Your color attribute
• Size: Your size attribute
• Gender: Your gender attribute
• Material: Your material attribute

Product Types - Configurable Products:

• Use Configurable/Simple Products: Only Linked Simple Products
• Use Parent URL for Simples: Yes, with Auto-Link
• Use Parent Image for Simples: Only if Empty
• Use Parent Data for Simples: Name, Description
• Use Fallback only on Non Visible Simples: Yes

Result: Each size/color variant exports separately, allowing AI to recommend specific options based on customer preferences.

Use case: Large catalog requiring frequent updates.

Configuration:

Automation:

• Enable Cron: Yes
• Frequency: Every 6 hours (or more frequent)

Filter Options:

• Use category filters to reduce feed size if not all products are relevant
• Enable stock filter to exclude unavailable items

Performance Tips:

• Filter aggressively—only include products you want AI platforms to recommend
• Exclude categories like "Parts" or "Accessories" that don't make sense in AI shopping contexts

Good titles help AI understand and present your products:

Example: "Nike Men's Air Max 90 Running Shoes - White"

Write descriptions that answer customer questions:

• What is this product?
• What are the key features?
• Who is it for?
• What materials/specifications?
• What's included?

AI platforms may extract information from descriptions to answer customer queries.

• Use high-quality images (minimum 500x500 pixels)
• Show the product clearly on a neutral background
• Include multiple angles if available
• Avoid watermarks or promotional text

Your return policy page should clearly state:

• Return window (number of days)
• Condition requirements
• Process for initiating returns
• Refund timeline

AI may reference this when customers ask about returns.

Include information about:

• Data collection practices
• How customer data is used
• Third-party sharing
• Customer rights

Cover:

• Purchase terms
• Shipping information
• Warranty details
• Dispute resolution