Overview

AI-Powered Product Recommendation System for Retailers

‍

Problems

This project addressed the challenges of providing personalised product recommendations in a large retail environment. Customers often seek assistance in selecting the right product, such as finding a suitable gift. This project aims to leverage AI to provide relevant product suggestions based on the product catalog, pricing, and customer needs.

‍

Solution

FlowFoundry's solution combined a Large Language Model (LLM) with the product catalog of a large retailer to provide intelligent product recommendations. The LLM is trained to understand the characteristics of products, pricing tiers, and customer preferences. It is capable of answering complex queries like, "Is there a mid-price gift I can get my mother for her birthday?" by analysing the product database and providing tailored suggestions that match the customer's criteria. This was packaged up into a mobile app so that it could be used on the shop floor of a retail outlet.

Key Aspects of the Project

1. Product Data Integration We integrate the retailer's product catalog, including details such as product descriptions, categories, and pricing. This comprehensive dataset allows the LLM to understand the full range of products available and their characteristics.
2. Model Training for Product Understanding The LLM is trained on product descriptions and customer behaviour data to develop an understanding of various product categories, pricing, and use cases. This training enables the model to make highly relevant product recommendations.
3. Personalized Recommendation Engine The system allows customers to ask questions in natural language, such as finding a gift or a specific product within a price range. The LLM processes these queries and provides recommendations based on product features, pricing, and suitability.
4. Natural Language Product Descriptions Once a recommendation is made, the LLM generates product descriptions and justifications for the choices, helping customers understand why a particular product is recommended. This feature enhances the shopping experience by providing transparent, informative answers.

‍