Create real business outcomes with AI/ML

Do you struggle with what digital means or do you struggle gaining real actionable insights from the massive amounts of data your company captures? Are you getting a real return on your investments in #AI/ML, #BigData, and/or #DataScience? Below is a quick example that highlights how these technical capabilities can benefit your business.

I recently built a neural network (aka #deeplearning) to predict customer churn for a bank. This was utilizing the archetypal dataset available on Kaggle.com. The insight gained was not necessarily that we could predict customer churn (which I can), but instead, WHO was churning. When I delved into the data to understand what were the most predictive features, I realized that the bank was losing customers that could be classified as their most valuable customers.

With this first level of insight, you immediately ask questions about your model. The first question I asked is "so what?" - meaning, there was nothing in this model that told me WHY these customers were leaving the bank. So you think how can I use data and AI/ML to answer why. First, since this dataset doesn't capture data to understand customer behaviors, this bank might undertake a customer segmentation study (aka unsupervised learning) to better understand HOW their customers interact with the bank, or how they would prefer to interact with the bank.

With this ensemble of predictive models, you now can make changes to predict if a customer could be at risk, and what you need to do to recruit and retain your best customers. You could even make a case that you can now make an informed decision about which customers your want to service. In this instance those informed decisions could result in new processes in how to take care of their best customers. Whether that is via digital interactions (mobile apps) or human interactions, your insights and resulting actions will be directed and you can feel confident that you are making the right changes to make the biggest impact to your business.

I would offer that this is one definition of digital transformation - using state of the art data modeling capabilities to allow you to move from feasibility to optimality, in your decision making and resultant business change efforts.