Our client, an ecommerce fashion retailer, saw 18% increase in order value and 21% drop in cart abandonment by using AI to provide personalized product recommendations and promotions for each individual shopper in realtime. An ensemble system with Azure-hosted LLAMA-2 language model was built to analyze customer's active browsing data and generate tailored suggestions in form of real-time popups. The conversational, context-aware promotional recommendations resonated strongly, driving revenue growth. AI personalization at scale powered by dedicated large language model instances transformed the customer experience of our client.