Machine Learning: Lessons for Banks From Self-Driving Cars

I read a McKinsey article the other day about 'cultural invisibility' - that tipping point when new technologies become so common they're taken for granted and become 'invisible'. Electricity, steam engines, and other twentieth century inventions typically seem to take about 80 years to make the transition. Computers haven't faded from view just yet, but are likely to do so by about 2040. And machine learning isn't expected to take much longer to recede into the background either.

Machine learning first came into its own in the late 90s. With the proliferation of unmanageable data volumes and complexity, data scientists realised that they could stop building finished rules-based models, and train computers to self-learn from data.

Most of us are already familiar with the concept of a self-driving car (machine learning in action). As it drives, it assesses risk much faster than human response times, taking evasive action to avoid collisions - even something happening 400 yards behind it. It remembers all the years of history and variations of conditions, such as tyre pressure, rain, ice, snow, specific junctions, and rush hour. And of course, at any time, the driver can just grab the wheel and take over.

Banks have the equivalent of their own 'self-driving' systems. Machine learning - in the form of complex predictive analytics, knowledge extraction, artificial intelligence and reasoning - is starting to perform tasks faster and more accurately than any human being. Just like specific driving conditions, these systems can instantly recall and process patterns faster and make accurate predictions or automated decisions for us.

In the same way humans will still be involved in the journeys of self-driving cars and take the wheel at any time, human bank staff will be involved with their systems – but much later in the process. I've outlined a few examples below:

Robo Investing – Machines investing in specified risk tolerance markets, taking evasive action when necessary (e.g. interest rates go up and oil prices drop).

Fraud Analysis - Track historical patterns, extract information and identify red flags that humans can't find in fraud analysis, money laundering and risk management. (The banking equivalent of avoiding accidents on the road).

Automating Payments - Machine learning automatically matches PO numbers with invoices, with staff getting involved once this matching has taken place.

Customer Service – Anticipating needs by intelligently tagging and clustering inbound social media posts, emails etc (triggering follow-up action if a customer gets married.)

At SAP, we're now embedding machine learning intelligence in business processes and making intelligent applications branded SAP Clea to make it easier to become an ' intelligent' business. I think moving forward, the accurate predictions, unprecedented insights and automated self-learning routines will all become culturally invisible.

In Europe, analysts say more than a dozen banks already profiting from machine learning by replacing their outdated statistical-modeling approaches. Some banks have seen 20% savings in capital expenditures, 20% increases in cash collections, as well as 20% declines in customer churn. Examples include building microtargeted models that more accurately forecast who will attrite or default on their loans, and how and when best to intervene.

Historically, reasoning has been a mostly human skill but it is now developing into a machine skill – and can't be ignored. I'll be discussing the key benefits, challenges and concerns around the spectrum of cognitive computing at the upcoming webinar, "Making Sense of Machine Learning In Financial Services" on 28 March 2017. If you'd like to get a better understanding into what these key game changers mean for you, you can register here: https://www.finextra.com/events/fastregister.aspx?eventid=157