Amazon Web Services has announced several major upgrades to its Amazon Personalize recommendation service, leveraging advancements in generative AI to take personalization to new heights.
The headlining capability is a new "Content Generator" feature powered by a foundation model that automatically generates natural language text to describe recommended items. This goes beyond generic phrases like "Frequently Bought Together" to craft one-of-a-kind titles and descriptions that match a user's unique interests.
For instance, an ecommerce site can automatically generate email subject lines specifically catering to what that shopper tends to purchase instead of sending blanket promotions. Similarly, a streaming platform can showcase personalized video collections with themes like "Adrenaline Rushes" or "Belly Laugh Bonanzas" based on an individual's watch history.
Amazon also unveiled Personalize on LangChain, enabling developers to integrate Amazon Personalize recommendations into their own generative AI applications built on the LangChain framework.
This allows builders to tap into Amazon's personalization algorithms while leveraging large language models within LangChain to create tailored responses. Use cases include crafting customized marketing messages in real-time, building AI-powered chatbots that give personalized product recommendations, and more.
Lastly, Amazon Personalize now returns select metadata fields along with recommendations as part of the inference response. This metadata, like item descriptions, genres, etc. can provide generative models additional context to inform even more relevant, targeted responses.
As explained by Fox's Head of Technology Daryl Bowden, these new capabilities will allow Amazon Personalize customers like Fox Sports to take personalization to new heights.
“Amazon Personalize has helped us achieve high levels of automation in content customization. For instance, FOX Sports experienced a 400% increase in viewership content starts post-event when applied. Now, we are augmenting generative AI with Amazon Bedrock to our pipeline in order to help our content editors generate themed collections.” – Daryl Bowden, EVP of Technology Platforms.
Apart from these features, Amazon Personalize facilitates a range of personalization experiences like tailored product and content recommendations, guidance on next actions, and user segmentation in targeted marketing campaigns. These can be achieved using recipes or algorithms prepared for specific use cases, requiring no ML expertise from the developers.