AI as banking’s WD-40: five examples of a frictionless future

Digital coins
(Image credit: Image Credit: Number1411 / Shutterstock)

The term ‘financial services’ conjures images of people in dark suits in tall buildings, when in reality it underpins the smooth running of your day-to-day life. Your daily customer experience in fact. It is fast moving towards a future of full integration to the point where it’s almost invisible.  

Consider how things have changed over the last 20 years. Chip-and-pin was only introduced in 2004, and contactless payments have only recently really been widely adopted, enabling an era of not only frictionless payments but also travel. Mobile banking has come on leaps and bounds, now using biometric authentication on your phone and looking to do away with the card reader machines of the past. Even your gaming experience has changed, allowing you to seamlessly buy custom skins on Fortnite in a couple of clicks. Just imagine what the next 20 years has in store!

For this ‘invisible’ vision to come true, financial services need to work on one main challenge: removing friction, be it from sales, complaints, or whatever transaction the customer wants. And the way this is achieved (at least in part) is through applied AI. This is no easy task. It takes a huge amount of data aggregation, modelling, anticipation, and risk management, not to mention a completely connected network of customer and internal touchpoints to be truly realised through applied AI. But the benefits could be in the orders of magnitude. Banking is in essence the business of managing risk, and spending on artificial intelligence does sound like a risky venture. Yet AI in fact presents a very simple business argument, especially to the ‘tier ones’ or ‘encumbants’: invest now, save later.  

To help paint a picture, here are five examples of where and how artificial intelligence and advanced analytics will help make future processes seamless:  

1. Mortgage selection, payment and settlement processes 

This area requires a fundamental shift in how we think of ‘the customer experience’ being delivered. The experience should be built around ‘having a new house’ (an exciting and joyous experience) versus the mortgage application process (a lengthy and painful process that needs to almost disappear). AI could, for example, allow banks to measure risk and calculate credit scores in advance so that the application process is already 99% complete before the customer is asked to input anything. Imagine logging onto your lender or brokerage and already having all of your provisional quotes lined up with no input from you!

2. Wealth maximisation and spending patterns

Everyone wants saving and growth to be as simple as possible. Making a customers’ wealth maximisation plan easier to set up, monitor and achieve would be another great form of applied AI. Surely an AI could optimise the combination of income, savings and taxation to give you the maximum wealth across an annualised period. The seamless customer experience would be that it would be easy to set up in line for your objective for the year, monitor what you need, and the banks would be able to offer advice throughout the year e.g. “We think your spending patterns could be improved, here’s how…” This feature could even become a point of competition between banks, with each touting their top-performing products to help customers hit their wealth maximisation goals.

3. Mortgage protection

Banks can come under pressure from regulators to accommodate customers that might default on their mortgages. However, doing-so without putting in preventative measures results in a bad time for all parties. So, banks need to know if and when this might happen. By training the AI (i.e. machine learning) using the previous behavioural patterns of the entire historical customer loan book – including those who have defaulted – banks are able to recognise similar behaviours well in advance of the event. The data here is more ‘operational’ than ‘transactional’, e.g. looks at how regular payments are, if ahead of time or late, how much etc. rather than based on current account spend.  This provides banks with the ability to engage and consult with the customer way ahead of time, and put in place solutions before the problem comes to a head. As an added bonus, they’ll also be able to demonstrate to the regulator that they’re doing so to the best of their abilities.

4. Complaints and customer handling

Using predictive analytics to identify when a customer is going to complain is invaluable. It augments the bank’s intelligence. Not only does it highlight internal issues, it allows them to fix or at least get in front of them before the complaint is logged, e.g. by refunding money prior to receiving an overdraft fee complaint and notifying the customer of this by text. This not only helps build a healthy customer relationship, it also saves money for the bank in the long term by mitigating account closures due to poor service. What’s more, analytics can also be used to monitor the emotional states of any callers, group them into personality and emotional states, and then offer guidance to the call handler on how best to steer the conversation.

5. Fraud detection

If you think about fraud detection, AI is already being largely employed to provide real time checks on fraud patterns. This used to be a lengthy process – taking many days, weeks, or even months through an application process. Now, using machine learning you can detect patterns very quickly. The machine’s decisions are based on historic patterns to detect when there is a likely fraud. And while it will never achieve 100% detection rate it affords a much quicker response to those likely outliers that need investigation. 

What’s more, once the machine has done so, you can also interrogate how the machine arrived at that decision (at least to a certain degree).  While banks are coming on leaps and bounds with this application of AI, there are still terribly ‘clunky’ practices such as video feeds, photocopying and so on involved in the process. As and when banks are able to generate a purely digital decision and arrive at that purely programmatically, that is when you will be able to interrogate each individual aspect of the decision making and arrive at ever more accurate detections. This all means that our money is in ever safer hands.

It’s important that financial services are able to look beyond the hype and recognise applied AI for the business opportunity it presents. These five examples alone warrant enough of a reason to at least think twice, and with the wealth of fintechs on the scene, artificial intelligence and our ‘invisible’ future is beginning to look a lot more realistic. 

Richard Hamerton-Stove, Principal at Capgemini Invent 

Richard Hamerton-Stove

Richard Hamerton-Stove is Prinicipal at Capgemini Invent. He is an experienced consultant and specialise in financial services. Leading an amazing practice full of wonderful data scientists who are consultants too. He deliveres large transformation programmes.