Introduction
Dynamics 365 Commerce AI recommendations provide optimized, machine-learned product lists that are available for retailers to implement across their businesses.
The Commerce solution can be used to show product recommendations on an e-Commerce website and Store Commerce devices. Using AI-powered recommendations helps improve product discovery and helps customers discover new content that they (the shopper) might not have realized they were interested in purchasing or would not have found otherwise.
Product recommendations can be integrated throughout the context of a shopper’s journey and can help shoppers quickly discover products that are relevant to them. This powerful reflection of user behavior and preference is beneficial to help shoppers find the right product in the right place in their shopping journey.
By using recommendations to enhance product discovery, retailers can create more conversion opportunities, help increase sales revenue, and even amplify customer satisfaction and retention. Cross-selling and upselling can also be used to assist customers in finding products that they did not originally intend to buy.
Retailers have discovered a multitude of application recommendations throughout their online stores, including:
- Landing page – Retailers can highlight products in the New, Best Selling, and Trending lists on their store’s landing page.
- Product details page – On this page, retailers can suggest additional items that are also likely to be purchased based on consumer behavior. These items appear in the People also like list.
- Transaction page or Checkout page – During checkout, retailers can suggest complementary items, to complete an order based on the existing contents of the customer’s cart. These items appear in the Frequently bought together list.
- Personalized recommendations – For signed-in customers, retailers can provide a personalized Picks for you list based on recent purchase history and can personalize existing list scenarios for each customer.
Product collection modules
Product collection modules help retailers build compelling shopping experiences by providing an intuitive visual interface that can be used to quickly compose product collections.
The following image shows the different types of product collections that are being used on an e-Commerce site.
The product collection module can be broken into four distinct types of collections:
- Editorial – A list of products that are manually curated and maintained in Commerce. This list includes products that are sorted by category, by their relation to other products, or are curated.
- Algorithmic – These lists contain items such as new, best-selling, or trending products.
- Machine learned, contextual recommendation – These lists are comprised of categories such as People also like and Frequently bought together.
- Picks for you – This list supports personalized results for signed-in customers only and are not available for guest users.
The following screenshot shows how to set up a new product list from a product list configuration that is being used on an e-Commerce site.
Recommendation service
The product recommendations service uses artificial intelligence (AI) and machine learning technologies by extracting the data from the Commerce operational database and then sending it to Azure Data Lake storage or an entity store. Then, this data is used to train recommendation models for the People also like, Frequently bought together, Picks for you, New, Best selling, and Trending lists.