Your AI Shopping Assistant is designed to help customers by reading your product information, answering questions, and making recommendations.
However, not all product data is meant for customers. Some product metafields may include internal information (such as profit margin), supplier details, or other attributes that should remain private. In these cases, you may prefer that Shoply does not use those metafields when answering customer questions.
Shoply uses AI to automatically determine which metafields can be used when answering customer questions. However, to give you full control, Shoply allows you to set blacklist (exclude) specific product metafields or entire metafield namespaces.
This guide explains why this is useful and how to configure it.
Go to the “Settings” section of the app’s admin console, and then scroll down to the “Advanced Settings” section. In this section, you can provide a JSON object to configure various settings.
Use these settings to prevent Shoply from using specific product metafields when answering customer questions.
blacklist_product_fieldsA list of exact metafield keys (field names) you want to exclude. The field name must match the metafield’s key exactly as it appears inside Shopify.
Example metafield keys:
start_dateproduct_typebackstock_numberupcshipping_lead_timebrandNote: if you don’t want to show vendor of the product, you need to add “brand” to the blacklist,
as Shoply uses the brand metafield to store vendor information for legacy reasons.
blacklist_namespacesA list of entire namespaces you want the AI to ignore.
Example:
RenterInfoIf a metafield belongs to a blacklisted namespace, all fields in that namespace are excluded automatically.
This is useful when a whole namespace is internal or sensitive.
(These help the AI give better answers.)
split_tag_to_metafield_byShopify tags are a convenient way to label products, but they’re stored as a flat list of strings and can’t be used as structured filters. This setting lets you convert key:value style tags into proper metafields that Shoply can index and use for filtering.
How it works:
Set split_tag_to_metafield_by to the delimiter character used in your tags. Shoply will split each tag on that character, treat the left side as the metafield key and the right side as the value, and index them accordingly.
For example, if your products have tags like:
material:aluminum
color:matte-black
compatible-with:iPhone 15Setting the delimiter to ":" will produce metafields:
| Key | Value |
|---|---|
material | aluminum |
color | matte-black |
compatible-with | iPhone 15 |
These indexed metafields can then be used as facet filters in Shoply AI Search — just like native Shopify metafields.
Tags that don’t contain the delimiter are ignored by this setting and remain as regular tags.
Example configuration:
{
"split_tag_to_metafield_by": ":"
}When to use this:
key:value tags for attributes (e.g., from a bulk import or third-party app)Use these settings to control what information is shown when products are displayed in the chat or search interface.
hide_discountWhen set to true, discount percentages (e.g., “-40%”) will be hidden from product displays. This is useful if you don’t want to emphasize discounts to customers.
hide_priceWhen set to true, all price information will be hidden, including original prices and discounted prices. Use this if you prefer customers to visit the product page to see pricing, or if your pricing strategy requires hiding prices in the chat/search interface.
hide_reviewWhen set to true, review ratings will not be shown on product displays. This is useful if you don’t want review scores to influence customer decisions in the chat/search interface.
Here is a fully valid example:
{
"blacklist_product_fields": [
"profit_margin",
"inventory_level",
"product_type",
"category_name"
],
"blacklist_namespaces": [
"SupplierInfo"
],
"split_tag_to_metafield_by": ":",
"hide_discount": true,
"hide_price": false,
"hide_review": false
}You can remove items you don’t need, or add your own metafield keys and namespaces. Set the hide_* options to true or false based on your preferences.