An Early Warning System Based On AI

Kirk says most advisors fall in one of two groups:

  1. Those who know about AI but don’t have a master data management strategy and/or approach to use their data.
  2. Those who are scared of the implications of AI. (e.g., “This is scary stuff. I don’t want to tell people they’re going to get cancer. How am I going to have this conversation?”)

Neither group is ready to embrace AI, so what changes should be made in the next couple of years?

Group 1 has to understand that without a good CRM or proper data management policies, you can forget about AI. Data has to be normalized so solutions like the one InterGen created can plug and chug. Clean data = better predictions = earlier warning of life events.

Group 2 has to realize that it is logically, morally and ethically right to warn a person if there is a chance of the individual getting sick. Granted, advisors are not doctors by any stretch of the imagination. Instead, it is the advisor’s duty to know if there’s a possibility clients need to spend money for an unexpected event.

“The challenge is making sure advisors and firms are ready for that change. I have data that can change their practices. I have data that can really help their clients,” says Kirk.

Not all use cases are medical-related. Kirk highlights an example of how their system can help in a use case for the mortgage industry. The U.S. mortgage industry is approximately $9.7 trillion and the default rate is a little over 3.2%. This means that defaults in the U.S. total up to around $310 billion. However, after much research there are only five main reasons that people default on their loans: 1) loss of job, 2) divorce, 3) death of spouse, 4) critical iIllness (Cancer), 5) Having to take care of a loved one.

Now as you can see, each of these are liife events, but if you were to apply InterGen Data’s AI solutions to mortgage clients, you would be able to quickly find correlations to other clients who have defaulted. This doesn’t necessarily mean that they “will” default, but the correlations could be used as an early warning system in conjunction with their established risk systems. If you have predictive criteria, you know when something’s going to happen before it happens and can be better prepared.

The Bottom Line

Deep learning algorithms were once mythical beasts. Now they’re being applied to everything from helping gamers find the quickest routes to beat video games, all sorts of prediction work and filtering like Google Translate does, and in the finance/investment world to pinpoint events that may trigger positive or negative outcomes. Nothing in the world is widely adopted right off the bat—it takes years of perfecting and molding the technology to provide the greatest benefit to users. AI and machine learning are no exception.