How banks can take advantage of machine learning?
It’s been almost 40 years since we know that the answer to the ultimate question is 42. Still we are asking less important questions but looking for more complicated answers.
As an innovation lab we are constantly asking questions, looking for problems and trying to pick the good ones. A good problem has 3 attributes:
- Someone cares about it.
- It could create value.
- It can be solved.
I believe the first and the second usually walk hand in hand and most people consider the third one as constant. Our main job is to find and push the technological limits and by result solve these new problems.
Nowadays the innovation has gone right to ludicrous speed. If you have an idea that you are not able to do it today, probably there will be a technological breakthrough tomorrow that will enable it.
Two of these great enabler technologies are machine learning and deep learning. These are currently at the peak on Gartner’s Hype Cycle and God knows, everybody is doing it. Why would we be different? Financial institutes have plenty of data, different kind of customers with various demands and not that hard to find realistic business models for the ideas.
Here are two examples. For free.
Financial advisory for everyone
People generate a lot of data with their financial activities. They make transfers, create deposits, use credit and debit cards at different merchants, invest, take loan and the list goes on.
Every interaction can be considered as a decision and the bank can analyse the past and give advice. They are only doing it mostly for premium customers because this is the only way to make this service profitable. With machine learning models the cost could be decreased at the level where the bank can give insights for every customer.
A few examples are account balance prediction, budget recommendation, investment portfolio management and automated daily savings. Actually these are not new ideas. Multiple fintech startups are doing these, but traditional banks are falling behind.
Increased profit by better customer insights
Customer experience is a good thing and can be beneficial on the long run but in the meantime something has to keep the lights on and this is efficient, immediate sales. This is the area where machine learning has been shining already. Ask Amazon for example!
There are tons of research and white papers about this topic. In a nutshell if you can successfully match one of your customer’s current attributes to one in the recent past, congratulations, you made a sale.
And the best of it that this actually works. You can start implementing it today and with a tiny luck get your first results the same day.
Yesterday was already too late
I hope these two short examples created a little bit of excitement and made you think about what you could do with this technology.
My advice is to start studying the field of machine learning and create your first experiments ASAP, because it can offer a lot of value not just for financial institutes but for everyone.
