• Evaluated past purchase behaviour of every customer through store, website POS and clickstream data and created a customer genome profile with knowledge accumulating with every interaction and activity
  • AI Picker engine selected the best performing algorithm for each customer based on context and data available from the pool of algorithms (session based, Hybrid filtering and deep learning). Used Multi Bandit Approach each favouring exploitation over exploration to different degrees to roll out recommendations
  • Filtered and aligned the products recommendation list to business goals and rules by increasing or lowering the weights of products


  • At granular category cohort level, the accuracy for Millennials were higher in Electronics and Gadget whereas accuracy was higher in everyday living category for customers with larger household.
  • Reduced the frequent grocery related and low-cost product recommendations with relevant and diverse recommendations
  • Tredence helped bridge the gap between stakeholders ensuring multiple integrations and adoption by providing right solutions and processes through consultative problem solving


  • The product personalizer resulted in 60% increase in purchase rates for both initial and returning sessions.