If you’re feeling overwhelmed ahead of Amazon Prime Day, a new AI feature from the e-commerce giant can help you score everything you want — even the items you don’t know you want yet. Amazon announced this week that it will give customers a more personalized shopping experience by using large language models (LLMs) to match individual preferences and provide more relevant product information.
The most useful, albeit subtle, change is the recommendation system. The list of products under “more like this” has always been based on the current product page you’re viewing, as well as your past searches and purchases. With the updated system, you’ll see ideas for things to buy based on specific details of your searches. Amazon’s examples describe how you might see gift suggestions based on a holiday you’ve previously bought gifts for, or recent deals on gear for a sport your history suggests you enjoy.
The actual descriptions of those recommended products now reflect your search history, too. A certain feature or adjective you include in your search query, such as vegan food or a certain color of furniture, translates to seeing those aspects highlighted in the product title. The goal is to help you make informed decisions without having to scroll through a bunch of details that don’t interest you, which saves valuable screen real estate on a mobile device. Combined, you should see better products and why they’re recommended immediately.
Prime plan
“If the primary LLM generates a product description that is too generic or fails to highlight key features unique to a specific customer, the evaluator LLM will flag the problem,” said Mihir Bhanot, director of personalization at Amazon. explained“This feedback loop allows the system to continually refine suggestions so that customers see the most accurate and informative product descriptions.”
This year, Amazon has been steadily rolling out new AI features for shoppers. The biggest addition is the Rufus chatbot, which uses AI to research products and recommend purchases through conversations rather than keyword searches. As with much of the rest of Amazon’s shopping experience, it will now include ads, but the general idea is still to tailor your shopping experience to your interests.
In the future, Amazon wants its tailored advice to pull from more than just your Amazon history. The company has plans for extensions to pull data from sources like Gmail and YouTube Music, which would further personalize the shopping experience. For now, expect your Prime Day to be a little more efficient, and the products you see recommended to be a little more like exactly what you didn’t know you were looking for.