BUSINESS SITUATION
- The client wanted to profile their customers and build an ML-based model to recommend allocation across their multiple micro-investment products of Invest Now, Invest for Retirement, Invest for your Kids, and others
- Existing allocation process was manual without any user profiling and for most of the cases, customers had to decide their investment split across various investment options
SGA APPROACH
- SGA profiled all the client’s customers basis their key features using a hybrid clustering technique that clearly defined the profiles aligning to business requirements
- Basis the user profiles, we designed the allocation engine that recommended the % split of investment amount across investment products
- The allocation recommendation was real-time, which was initiated during the user registration process for new users and for old users, it was provided under the recommendation section in the app
- We deployed the model using MLOps and built an end-to-end data pipeline for model serving and feedback collection from the customer for model refinement
ENGAGEMENT
The client is a US-based financial technology and financial services company that specializes in micro-investing and Robo-investing
BENEFITS AND OUTCOMES
- The data pipeline was created to fetch data from multiple sources to create the master dataset for feature creation and user profiling
- We created a feature store after looking at 800+ metrics currently captured by the client app and filtered out the key metrics for translating into features
- The automated and intelligent allocation enabled higher customer satisfaction, owing to ease of usage
- Auto-allocation helped in higher investment and usage because of the customized allocation, reducing the hassle of allocation decisions by customers
KEY TAKEAWAYS
- The profiling helped in providing customed allocation across the customer base for both new and existing ones
- The investment across products increased as most of the customers followed the recommended investment split
