Personal Shopper Agent

This agent provides personalized shopping options by analyzing a customer’s shopping history and purchase pattern. It suggests complementary products the customer is likely to need or want next, creating a more intuitive and relevant shopping experience. Rather than relying only on historical data matching, it uses the contextual understanding of Large Language Models to present a wider, more logical range of product suggestions.

The Complementary Products Agent uses an LLM to generate complementary product categories, then runs an embedding search to find matching products within those categories. This approach supports common-sense product pairings and wider matches, even without extensive historical data across all customers. Retailers can also adapt it to promote specific product categories during sales events or seasonal campaigns. The intended impact is higher average order value, incremental sales from cross-selling, and greater customer loyalty and engagement.