Embedded Files
Introduction
Catalog training allows your AI agent to understand your products and use them during conversations.
This data source is used by Shopping Assistant agents and enables the agent to answer product-related questions, assist with discovery, and support ecommerce journeys using your actual inventory.
Catalog training helps turn your AI agent into a knowledgeable sales assistant rather than a generic chatbot.
What Catalog Training Does
When you train an agent using a product catalog:
Product data is stored as structured knowledge
The agent understands products, variants, and attributes
Conversations are grounded in real inventory data
Catalog training allows the agent to:
Answer product-specific questions accurately
Assist users while browsing or comparing products
Reference real products instead of generic suggestions
Who Can Use Catalog Training
Catalog training is available for:
Shopping Assistant agents only
Other agent roles (Customer Service or Employee Support) do not use catalog data.
Where to Manage Catalog Training
To manage catalog training:
Open your Shopping Assistant agent from the My Agents page
Go to Knowledge Base
Select Catalog
This section displays all products currently available to the agent
Ways to Train the Agent with Your Products
There are two supported ways to train a Shopping Assistant agent using your product data. The method you choose depends on your ecommerce setup.
1. Train Products via APIs
If your store runs on a custom website or non-Shopify platform, you can train the agent by connecting your product APIs.
Using APIs allows you to:
Add your product catalog programmatically
Keep product data updated in real time or at scheduled intervals
Sync pricing, availability, and inventory directly from your backend
This method is ideal for:
Custom ecommerce websites
Headless commerce setups
Businesses with frequently changing inventories
Once connected, the agent automatically uses this product data during conversations.
2. Train Products via Shopify (Automatic Sync)
If your store is built on Shopify, product training happens automatically.
When your Shopify store is connected:
Products are fetched directly from Shopify
Catalog data stays in sync automatically
Variants, pricing, availability, and tags are included
No manual upload or API setup is required
This is the fastest and simplest way to train a Shopping Assistant for Shopify stores.
Choosing the Right Method
Use Shopify sync if your store is on Shopify
Use API-based training if your store is custom or non-Shopify
Both approaches result in the same outcome: a Shopping Assistant that understands your products and can use them during conversations.
What Product Information Is Used
Catalog training typically includes:
Product name
Product description
Variants
Pricing
Availability
Tags and attributes
The quality and completeness of this data directly affect response accuracy.
Importance of Product Tags
Product tags play an important role in catalog training.
Tags help the agent:
Group related products
Understand categories and use cases
Improve relevance during recommendations
For best results:
Use clear and meaningful tags
Avoid inconsistent or duplicate tagging
Keep tags aligned with how customers search
Well-tagged products significantly improve shopping conversations.
Managing the Product Catalog
From the Catalog section, you can:
View all products available to the agent
Monitor catalog status
Remove products if needed
For Shopify and API-based catalogs:
Updates happen automatically
For CSV-based or one-time imports (if used):
Re-upload is required to apply changes
Combining Catalog Training with Other Data Sources
Catalog training works best when combined with:
Website training for store pages and policies
Document training for manuals or detailed specs
Text training for rules, exceptions, and sales guidance
Catalog data explains what you sell, while other data sources explain how you sell it.
Best Practices for Catalog Training
Keep product data clean and up to date
Use consistent naming and tagging
Remove discontinued products
Test product-related questions regularly
Pair catalog training with clear AI persona settings
A well-maintained catalog leads to better recommendations and smoother conversations.
Common Mistakes to Avoid
Missing or poorly defined product tags
Outdated pricing or availability
Duplicate products across sources
Expecting recommendations without proper catalog setup
If a product isn’t in the catalog, the agent can’t recommend it.
How to Validate Catalog Training
After catalog setup:
Open Test AI Agent
Ask product-related questions
Check whether responses reference correct products
Validate variants, pricing, and availability
Use Instant Retrain if refinements are needed
Testing ensures the agent uses catalog data correctly.
When to Review or Update Catalog Training
You should review catalog training when:
New products are added
Existing products are updated
Products are discontinued
Tags or attributes change
Keeping the catalog current keeps recommendations accurate.
What’s Next?
Once catalog training is complete, you can fine-tune how products are recommended to users.
Next User Guide: Product Recommendation
Page updated
Report abuse