When we think of compliance and regulations, there is typically a tendency to equate that with costs. There is absolutely no doubt that complying with regulations requires significant resources. By some estimates compliance and regulation costs the banking industry $270 billion a year and that it accounts for 10 per cent of operating costs. Compliance lapses can cost dearly as well. In 2012 HSBC agreed to pay a $1.92 billion in fines to U.S. authorities for compliance lapses which included AML issue.

But here is the thing, compliance lapses due to a lack of reliable compliance processes and mechanisms can potentially lead to loss of trust among consumers and hence, loss of value in the long term. We have seen data from many surveys about how Millennials do not trust their banks (this is due to many reasons of course—but compliance lapses do not help).

I would argue that we should treat compliance not just as a cost, but as something that requires investments in the form of latest technology such as AI, because its value is not appreciated until bad outcomes lead to loss of trust. And trust is something that is at the core of the banking business model and its most important asset; all of us depend on our banks to keep our money safe.

This is also perhaps an important reason for combining resources across the banking industry, and combine data to make the AI algorithms much more reliable. The effectiveness of AI and machine learning is heavily dependent on the data that is used to train the algorithms.

Combating fraud is clearly one of the most effective use cases of AI and machine learning—not only to cut costs but also to increase long term value in the form of strengthening trust among consumers.