Tips and Tutorials
July 7, 2023

What Is AI-Powered Recommendation And Best Practices For Shopify Merchants


We’ve seen the power of applying artificial intelligence (AI) to the online retail industry. With the increasing use of AI for small and large businesses, it’s important to understand and know how to optimize it for more tremendous online success. In today's article, we’ll talk about one of the most employed tools in eCommerce - AI-power recommendations - and see how Shopify merchants can apply some best practices to their stores. 

What is AI-Powered Recommendation? 

Before jumping to the actual case studies, let’s check if you have the right idea of what AI-based suggestions are. 

Everyone is talking about AI. Artificial intelligence (AI) is the ability of a computer or machine to think and learn. It is a branch of computer science that focuses on developing intelligent machines that function like humans. Hence, it helps us automate complex processes, improve decision-making, and identify actionable insights in big data. 

What about product recommendations? As its name implies, product recommendations intelligently suggest products based on aggregating store shopping data and the ability to pick certain items for promotions manually. 

So when we talk about AI-based product recommendations, we’re referring to recommendations automatically generated by AI models. These suggestions are based on products commonly purchased together as well as items with similar product data, and become more accurate over time as more order and product information become available.

How Does AI-Powered Recommendation Work? 

AI-powered recommendations work in complex ways. They use a recommendation system to filter and determine how someone would rate or favor a particular item. It makes use of a series of algorithms, analyzes data, and employs artificial intelligence (AI) to make recommendations. A recommendation system, in general, is an algorithm that proposes appropriate products to customers. 

Are you still confused? We list some popular recommendation types you would be familiar with once you see their names. 

Frequently bought together is one of the most used suggestion types on digital stores (Source: Boost AI Search & Discovery)

Apart from the frequently bought together widget, you will often see some blocks recommending related items, newest arrivals, trending products, bestsellers, most viewed, recently viewed, or handpicked products from the homepage to the checkout page. 

Why Do Product Recommendations Matter? 

The market-leading online retailers like Amazon are fluent in how AI-powered recommendations benefit their business from the bottom line. If store owners execute them effectively, these blocks will significantly benefit both e-shoppers and online sales.  

  1. They streamline product discovery by showing hyper-relevant and personalized recommendations. Putting the appropriate product in front of the right consumer at the right time is key to boosting customer satisfaction. 
  2. They build trust through smart and tailored experiences. Consumers would see that you worked hard to understand their demands and improve the experience impressively, which positions your brand as a credible retailer. 
  3. They maximize average order value (AOV). This is especially true if online merchants use personalized cross-selling and upselling suggestions at any touchpoints to urge shoppers to add and purchase more items. 
  4. They boost the conversion rate up to x5.5 higher. Targeted recommendation widgets enhance the overall shopping journey, turning browsers into paying customers. Later, they could even turn into brand advocates.

Best Practices To Implement AI-Powered Recommendations On Shopify Stores

Strategically managing product suggestions is an ongoing task that needs regular optimization. Whether you're a novice or a seasoned user in this field, let’s check out the following approaches when applying intelligent product suggestions. 

The good news is, product recommendations can be used almost anywhere on your online store to engage customers and urge them to purchase. You can place the widget on the homepage, collection, product, and cart pages. The secret is to choose the type that makes sense to the touchpoints during a shopping journey. The following list nicely gives you an idea of which recommendation types should go with different pages on your stores. 


  • Newest arrivals
  • Trending products
  • Bestsellers
  • Hand-picked products

Collection Page

  • Trending products
  • Bestsellers
  • Most viewed

Product Page

  • Frequently bought together
  • Related items
  • Most viewed
  • Recently viewed
  • Hand-picked products

Cart Page

  • Frequently bought together
  • Related items
  • Trending products
  • Bestsellers
  • Recently viewed
  • Hand-picked products

Now, we'll walk you through tried-and-true tips on implementing AI-based product recommendations on each page with actual use cases. 

On Homepage

Your homepage is a golden real estate that makes the first impression on prospects. Every visitor will visit your homepage first and scroll to find what they are looking for. Hence, you should quickly capture their attention by showing products that match their interests. As mentioned above, the home page is ideal for placing recommendation blocks such as Bestsellers, New Arrivals, and Trending Products. 

Best-selling items are a great starting point on the home page. You can add up to 3 recommendation widgets to impress the customers on this page. (Source: SandDollar)

SandDollar is a brand selling beachwear in Dubai. When a shopper lands on their homepage, they will see an eye-catching section called Best Seller of the Week, including a necklace, beach bag, sandals, and dress. Did you notice that if you buy all four items, it will give you a complete look? 

On A Collection Page

We all know collection pages are a great asset in driving conversions and enhancing product awareness for your e-shop. It allows you to organize products in a structured way and then helps shoppers browse and find items much faster.

On the collection page for eye makeup products, HC Forniture has space to cross-sells its trending products such as hair dye, nail polishes, or shampoos.  (Source: HC Forniture

Think about setting up 1 or 2 recommendation widgets at the end of the collection page. This setting should go with pagination or pages with a ‘View more' button. The popular items are based on the view or purchase events in the recent 1 to 7 days. 

However, don't leave the recommendation widget at the end if you have an infinity-scrolling page. 

On A Product Page

The most used recommendation widgets on product pages are related items and frequently bought together. Yes Bebe places these recommendations below the main product description. You can also set up the Recently Viewed widget on the product page, right above the footer.

It's good to go with 2 - 3, or a maximum of 4 suggestion blocks on the product page to avoid distracting shoppers. (Source: Yes Bebe)

On the page for a  travel cup, Yes Bebe showcases 5 other items including a bubble wand, mirror ball experiments, a towel, and bubble making set. All these products and  the travel cup have one thing in common - they are appropriate for family outdoor activities like spending quality time together on a picnic. 

On Cart Page 

It's suitable to place 1- 2 widgets on the cart page. Showing recently viewed products like Kanopi would be a wise choice to encourage customers to buy more and easily go back to products they were interested in. (Source: Kanopi)

Before shoppers go to checkout and purchase, the cart page is the last time for you to convince them to buy more items. The most used recommendation types on the basket page are best-selling and recently viewed products (this is based on their browsing history). The ideal placement for those widgets is below the cart information but above the footer.

Wrapping Up

AI-powered product recommendations are essential. They quickly and easily sharpen the relevancy and boost customer engagement, conversions, click-throughs, average order value, loyalty, and retention. 

You can now launch effective campaigns after understanding the many product suggestions, which types fit which area on-site, and the dos and don'ts.

Always keep in mind that your business is unique. You have a distinct brand, personality, audience, and products. Thus, your product recommendations should harmonize with all these, which you can customize, optimize, and test.

CiCi Nguyen

CiCi Nguyen is a Content Specialist at Boost Commerce. She loves doing research and writing to share helpful information with readers. Outside of work, she is a food lover and passionate about cooking and reading comics.