What in the world is "Machine Learning"?
Artificial Intelligence. Machine Learning. These two phrases have to be the most misused/abused phrases in the last couple of years. There’s no company worth its salt which doesn’t claim to be using “AI+ML” - even though the phrase itself is technically incorrect.
So, why should you pay attention when we tell you about our Machine Learning initiatives at Kenko? Because it’s meant to make life better for you. It's not just a jargon for us.
First, let’s get some basics out of the way and (hopefully) shatter some myths. Machine Learning is a SUBSET of Artificial Intelligence and it is NOT new. The term itself was coined 60 years ago, way back in 1959! Machine learning has its ancestry in heuristics and predictive modeling. Heuristics involve telling a computer how to complete a particular task in the most efficient way. For instance, the growth rate of the population can be predicted using heuristic models.
Machine Learning, on the other hand, changes the paradigm. It involves TEACHING the computer how to solve a problem by showing thousands of solutions to similar problems.
At Kenko, we believe that something like Machine Learning is not one for writing in advertising copy but an efficient tool to make the experience better across our offerings.
For instance, in our Kenko scoring questionnaire, we gather social and demographic data from prospective customers. Our ML algorithms then use these to predict the medical profiles of individuals in specific groups. Over time, we will not only be able to limit the questions we ask but we will also be able to provide more targeted treatments for conditions such as diabetes.
At the same time, we are using ML to limit or remove the mechanical, non-value added activity from our apps and websites such as having to press a ‘submit’ button. Sounds a little flaky, but if we can crack it, it'll be a huge step up in user experience.
We are also using intelligent engines to ask users to give us their KYC through a video selfie where they state their name, age, and location, etc. We use natural language engines to detect what the person is saying. In the (near) future, we plan to use facial topography and voice patterns to root out fraud. Most financial services companies have frauds ranging from 8-12%, with health insurance sector closing in on the 12% end of the spectrum . By using ML, we will be able to significantly minimize fraud which should result in direct savings for customers because lesser frauds would mean that we will be able to keep our prices super low. After all, why should you pay higher prices because someone else is trying to game the system?
Machine Learning has the potential to deliver better outcomes for both our customers and us. However, like everyone else, we need to work hard and long to make sure we're not just using it as fancy jargon. We've taken baby steps in this direction and look forward to your encouragement for the countless more to come.