04. 06. 2018
“The odds don’t seem to be in favour of traditional banks”
“Banks still lag behind in adopting artificial intelligence and machine learning because changing their internal systems takes way too long and they are fishing in a small talent pool,” says Remco Veenenberg, W.UP’s very own international sales evangelist. He’s recently joined us from City.AI, an international network of AI communities.
Originally from the Amsterdam startup scene, he’s now based in Budapest, a city he finds fascinating because of its local talent and opportunities to generate impact. His own community, Budapest AI, has set out to strengthen the city’s tech ecosystem by organizing knowledge exchange events and smoothing the way for business matchmaking. We sat down with him to talk about how data-driven banking could be a means of survival for banks – and how it could have saved him from a dreadful experience on his last trip abroad.
AI talent pool is extremely small
You’ve spent the last two years taking a deep dive into AI and ML by setting up more than 40 applied AI communities worldwide. When it comes to incumbent banks, how do you see AI and ML’s role and their level of adoption today?
In the past couple of years, basic ML and AI technologies have been widely adopted by big tech giants, fintechs and many others. Banks, however, are far behind. In the US, huge players, such as JPMorgan and Wells Fargo, are currently using and developing robust AI technologies. It might sound odd, but what I found while working for City.AI is that the global ML and AI talent pool is extremely small. It’s just so difficult to find and keep talent. Companies such as Google, Amazon and Facebook are developing solid AI platforms and they’re much more agile and modern, plus they usually offer better salaries and perks to employees.
AI has been a buzzword for traditional financial institutions that want to go digital. Which areas or functions do you think traditional banks are making the biggest progress in and where are they lagging behind?
The ‘AI hype’ has lead to pretty unrealistic expectations among decision-makers, and not just in the financial industry. Whether banks are making real progress in leveraging AI is highly doubtful. Most financial institutions are focusing on improving traditional bank functions like money management, lending, credit cards and insurance. Data-driven decision-making might be gaining traction but we can’t call it true AI or even necessary to reach these companies’ goals. But as soon as these solutions prove successful, banks are hasty to catch up.
One of the main areas where data-driven decision-making and automatization are a hit is customer support. Think chatbots, even if their usefulness has come under fire recently. Other areas are fraud management, personalized marketing, wealth management and trading execution on the private banking side. Most large traditional banks are lagging behind in almost every area. This is mainly because the time they need to change their internal systems is far too long and the process is rather bureaucratic.
Horrible banking experiences
Making better use of the wealth of existing customer data has been a key goal for banks looking to boost digital sales. How can AI and ML help these institutions achieve this?
Banks should focus on making their customer experience better. With almost disturbing regularity, I hear friends and peers complain about horrible banking experiences. Today’s customers are fed up with hidden fees, long queues, unresponsive customer support or hefty fees for exchanging currencies. To make the experience better, banks will have to start mining out various sources of data to be able to make tailor-made offers exactly when customers need them.
Let me give you a personal example. A while ago I was travelling to Tel Aviv, only to find out my card had been blocked. My bank could have easily predicted my trip and warned me to change my global security settings. But they didn’t so I was left without any cash. I had no choice but to make an expensive credit card withdrawal. An excellent solution could have been W.UP’s Sales.UP. Besides banking data, it uses different sources of information, such as geolocation.
Or take the automatization of new service purchases. It’s a critical area where going digital could mean a world of difference. Given all the regulatory issues and changes banks face, however, the current experience at most banks is still outright terrible. They should really focus on improving this. If AI and ML can help them, they have two choices: get the proper technology through a long and costly internal development process or buy it from the right external vendor.
Challenger banks pose a serious threat
What can traditional banks learn from digital-only challengers and big tech firms about the use of AI and ML? How could incumbents benefit from teaming up with third parties and fintech firms in this area?
Challenger banks such as Revolut, Monzo, Starling and N26 have solved the biggest problem people have with banks: the overall horrible experience. Offering a highly optimized user experience, one-tap banking, no hidden fees, instant service purchases and extremely cheap currency exchange, challenger banks pose a serious threat to traditional banks.
I left my own Dutch bank a while ago because every interaction took ages and they charged me ridiculous amounts for exchanging euros to forints. What does this tell banks? It’s time they beefed up their customer experience. They’ve been in the industry longer, they’re bigger in size, and they have more experience and capital than challengers. Not to mention that most customers won’t leave their bank unless there is a very good reason or a much better alternative. With tech giants lurking around the corner and Amazon getting into retail banking, the odds don’t seem to be in favour of traditional banks.
Given the time you need to change entire internal operations, I believe the winning strategy for banks is to team up with the right fintech that already implemented AI in their models and have relevant expertise. And a top data science team doesn’t hurt either.