Skip to main content

How to set up rule-based recommendation

In this guide, we will walk you through the process of setting up Rule-based Boost AI Search & Discovery Recommendation widgets

Thomas Ta avatar
Written by Thomas Ta
Updated yesterday

⚠️ This feature only works with Boost AI Search & Discovery V2 and above. To find out which version of our app you are using, please follow this article.


Understanding rule-based recommendation

Rule-based recommendations are generated by the rules set up by your store. The recommended products will be based on product attributes such as collection, vendor, tag, metafield, and product type.

Two types of recommendation widgets can use rule-based recommendations Frequently bought together and Related items.


Set up rule-based recommendation

As a basic flow to set up a recommendation widget, to create a rule-based widget in the app’s dashboard:

  1. Go to Merchandising > Recommendation > Widgets.

  2. Click Add new widget.

  3. On the Recommendation Type section, select Product page or Cart page > select Frequently bought together or Related items > click Next.

⚠️ Currently, Rule-based recommendation is only available for Frequently Bought Together and Related items recommendation types.

4. On Recommended product settings, select Rule-based recommendation.

5. Toggle the desired rules to display recommended products. The recommended products will have the same attribute(s) as the source product.

  • Same collection (once enabled, you can select which collection to ignore)

  • Same product category

  • Same product type

  • Same vendor

  • Same tags (once enabled, you can select which tags to ignore)

  • Same product metafield value 1-3

You can also rearrange the rules list by simple drag-and-drop method:

⚠️ The maximum number of rules that can be enabled at the same time is 4.

💡

  • When multiple rules are enabled, the model uses OR logic—any product matching at least one rule can be recommended.

  • More matches = higher rank. Products that satisfy more of your rules rank above those that match fewer.

  • Set priority by drag‑and‑drop. Higher‑priority rules surface their matches before lower‑priority rules; the top rule controls the first position on the storefront.

  • Example:

You enable two rules: Same collection (1st) and Same vendor (2nd).

  • Recommended products include items that match either rule.

  • Items matching both rules rank highest.

  • Next, products from the same collection appear, followed by products from the same vendor.

Use ignored fields to exclude specific collections, tags, or metafields from the model. These values are removed from the input.

  • Example: A source product has tags A, B, C. If you ignore tag B, the model treats the product as having only A and C and recommends items that match A and C.6. Click Next.


Tutorial Video

Watch our tutorial video to learn how to set up a Product Recommendation Widget:


Feel free to reach out to our dedicated support team via chat if you have any questions or require additional assistance.

Did this answer your question?