H2OWorld

Description:

Transamerica built a product recommendation system that can be leveraged across multiple distribution channels to recommend products to serve customer needs and reduce complexity. This session highlights machine learning technology and models, big data architecture that enabled us to build product recommender platform.

Abstract:

Transamerica sells wide array of products in Insurance, Investment and retirement space. Due to complexity and variety of our products, selecting a right product that serves customer needs becomes challenging. We built a recommendation engine to guide the customer and simplify customer experience. Recommendation engine is also leveraged by marketing and customer service teams for cross sell and upsell offers. 

We developed a customer data hub providing 360 degree view of the customer on Cloudera’s Hadoop platform with data feeding from all our administration systems. Recommendation models were built using H2O machine learning platform on our customer data hub.

In this presentation, we will focus specifically on our machine learning models and data challenges engineering team faced. The presentation will also highlight the architecture of the customer data hub.