Basic flow to set up a Boost Recommendation widget

📖 Boost AI Search & Discovery's Recommendation widgets are a powerful tool to enhance your online store's performance by suggesting relevant products to your customers. This step-by-step guide will walk you through the basic flow to set up a Boost recommendation widget and start driving conversions and sales.

On this page

⚠ This feature only works with Boost AI Search & Discovery V3 and above. To find out which version of our app you are using, please follow this article.

Definition of our recommendation types

In preparation for configuring a recommendation widget, it's essential to familiarize yourself with our distinct classification of recommendation types:

Recommendation Type Definition
Frequently Bought Together Recommends products that other customers have added to their carts along with the source product. This type suggests items that are commonly purchased together, enhancing the overall shopping experience.
Related Items Provides suggestions for alternative and complementary products that share common attributes with the source product. Complementary recommendations help customers discover add-ons or accessories, while alternative product suggestions offer similar yet distinct options.
Newest Arrivals Recommends recently added products that have been published to your store. This type keeps customers updated with the latest offerings and encourages them to explore new arrivals.
Trending Products Highlights popular products that have been viewed or purchased frequently within the past 7 days. This type capitalizes on current trends and customer preferences.
Bestsellers Recommends top-selling items based on the frequency of customer purchases over the last 30 days. This type showcases products that are proven favorites among your customers.
Most Viewed Suggests popular items based on how frequently customers have viewed them within the last 30 days. This type helps highlight products that have captured customers' attention.
Recently Viewed Recommends products based on customers' recent browsing history. This type reminds customers of items they've previously shown interest in.
Hand-Picked Products Provides manual recommendations curated by the merchant. This type allows you to showcase products that you've personally selected to enhance the customer experience.
Recently Purchased Showcases items that have been bought recently by other customers. This not only provides social proof, enhancing trust and credibility, but also keeps your product offerings dynamic and fresh. It taps into the power of customer behavior, encouraging visitors to explore what others are buying. By highlighting recent purchases, this feature can effectively drive attention to popular items and create a sense of urgency, potentially boosting sales and customer engagement.

Product recommendation requirements

To ensure the effective display of recommended products, it is imperative to adhere to the following criteria:

  • Availability: Recommended products must be in stock. There are two scenarios for sold-out products:
    • The products belong to non-fulfillment stores.

  • Pricing: Products should have a price higher than $0.00. Recommendations will not be made for products with a price of $0 or products without a listed price.
  • Publication: Recommended products must be published on your Online Store sales channel.

It is important to note that these criteria are applicable at the product level and not the variant level:

  • If a product boasts multiple variants, even if only one variant has a quantity of ≤ 0, the product will still be recommended, as long as it remains in stock.
  • Conversely, if a product features several variants and at least one of those variants has a price of ≤ 0, the product will not be recommended. This measure ensures that shoppers on the storefront will not encounter products listed with a starting price of 0.

These stipulations are designed to facilitate customers in discovering and purchasing products that are currently available.

Create a recommendation widget

To craft a recommendation widget that seamlessly integrates into your online store, meticulously follow these steps:

Select page type

  1. Open Boost AI Search & Discovery, from our app dashboard > Recommendation > Recommendation widgets.

  1. Click Add new widget.

Select Recommendation type

  1. Select the page type you'd like to add your new Recommendation widget (Homepage, Product page, Collection page & Cart page).

  1. Keep in mind that each chosen page will correspondingly entail a distinct type of recommendation widget.
Recommendation type
Collection page
Product page
Cart page
Frequently bought together

✔️ ✔️
Related items

✔️ ✔️
Newest arrivals
✔️ ✔️ ✔️ ✔️
Trending products
✔️ ✔️ ✔️ ✔️
✔️ ✔️ ✔️ ✔️
Most viewed
✔️ ✔️ ✔️ ✔️
Recently viewed
✔️ ✔️ ✔️ ✔️
Hand-picked products

Recently purchased ✔️ ✔️ ✔️ ✔️

Your choice of page type lays the foundation for the subsequent customization steps, allowing you to tailor your recommendation widget to suit the specific needs of your online store.

  1. Click Next.

Set up General settings for the widget type

For those seeking more tailored recommendations, we also offer specialized types with advanced setups, empowering merchants to cater to their distinct needs:

  • Frequently Bought Together and Related Items recommendations can be generated through either AI-based algorithms or Rule-based criteria, which encompass factors like collection, vendor, tag, metafield, and product type.
AI-based recommendation
Rule-based recommendation
  • The Hand-picked Products widget enables you to curate recommendations manually, utilizing the Manual Settings feature. This grants you full control over the selection process. If you haven't set up for this widget, follow this article.

  • Want to stay ahead of the trends? The Trending Products widget can be configured based on customer views and purchase events throughout your store. You can even define the time frame, ranging from 1 to 7 days, for calculating these trends.

Widget design

Currently, we offer support for two distinct widget layouts across four primary pages (with the Bundle layout, exclusively available on the Product page):

Model | Type Carousel (Slider) Bundle
Frequently bought together
✔️ ✔️(only for product page)
Related items
✔️ ✔️(only for product page)
Newest arrivals
Trending products
Most viewed
Recently viewed
Hand-picked products
Recently Purchased ✔️

To learn how to design a widget, please visit this article.

Edit recommendation widget

To edit a recommendation widget, go to Recommendation > Manage recommendation widgets.

  1. On the widget list, based on the Widget ID and Name to select the recommendation you want to edit by clicking on the widget's Name or button “⋯” from the Actions column.

  2. On the Setup widget screen, search for and manage other widgets from the taskbar:

  • Live: all your published Widget(s) will be listed here.
  • Draft: all your unfinished setup Widget(s) will be listed here.
  • Archive: all your archived Widget(s) will be listed here.
  • Search bar: search for a specific widget with its Name/Widget ID.
  • Filter: filtering and customizing your widgets list.

  • Refresh: reload your widgets manager list for the latest updated one.


Actions Description
Edit This action allows for modifications to existing content or settings. It provides the flexibility to change or update information, ensuring that data remains current and accurate.
Duplicate This creates an exact copy of the selected item. It's useful for replicating settings or content without having to recreate them from scratch, saving time and effort.
Pause Temporarily halts an ongoing widget campaign. This is beneficial for stopping activities without losing their settings or data, allowing for later resumption.
Archive & Restore Archiving moves items out of active use, decluttering the workspace without permanent deletion. Restoring brings archived items back into active use, making it easy to retrieve and reuse previously stored data or settings.
Delete This action permanently removes an item from the system. It's used to clear unwanted or unnecessary data, ensuring that only relevant and current content is retained. However, caution should be exercised as deletion is often irreversible.


This widget is published or enabled on the storefront.
This widget has NOT been published on the storefront yet.
This widget is archived when users no longer want to show on the Drafts list.
In case the widget has already been created or published on the live theme but the store is not subscribing R add-on or advanced packages.

Tutorial Video

Watch our tutorial video to learn how to set up your Boost Recommendation widget

If you need any further assistance, please do not hesitate to contact us at