AI Product Recommendations for Globus with Vertex AI
Globus engaged KPS to help prototype and build an AI based product recommendation engine with product master data joined with purchase data coming from SAP Customer Activity Repository (CAR) export (through SAP Business Warehouse).
Contact KPS
The challenge
Globus faced a significant challenge with product recommendations that lacked personalisation, particularly for unknown or unregistered customers. The existing recommendation system was not tailored to individual customer preferences, resulting in generic product suggestions that did not enhance the customer shopping experience or drive optimal sales performance.
The solution
To tackle this challenge, KPS utilised Google Cloud Vertex AI to develop AI-based product recommendations using SAP CAR point-of-sale data. Our approach involved importing raw SAP CAR data into Google Cloud, transforming it, and persisting it in Google BigQuery. We trained an AI recommendation model that integrated product and purchase data, achieving a confidence level of at least 80%.
The result
We consumed successfully 95,000 user events and signals as well as 170,000 catalogue data. In under 4 weeks we delivered this project and received a prediction confidence level of over 80% with the first shot.
About Globus Holding GmbH & Co. KG
Globus, a family owned and operated food and non-food Retailer, operates 95 stores in Germany, Czech and Russia. KPS is delighted to accompany Globus on its transformation journey since 2018 and support them on this new and innovative Project.
- Industry: Retail & Wholesale
- Primary project location: Germany
