Jim Marous’s latest digital banking report is out and, well, it ain’t pretty. “While there is an almost universal awareness of what needs to be done to digitally transform financial institutions, research from the Digital Banking Report shows that the progress is still slow to non-existent at most organizations. This is a recipe for failure in today’s banking ecosystem,” the financial industry strategist sums up. Sounds fair, considering the report’s key findings, most importantly that retail banking transformation is stuck in infancy stage and the majority of banks still haven’t leveraged advanced technologies, such as robotics, design thinking or artificial intelligence.
Here’s our problem with that: when it comes to going digital, half-baked efforts won’t get banks anywhere. “Do… or do not. There is no try,” Yoda instructs Luke in The Empire Strikes Back in a lesson that rings just as true for banks today.
“Before the digital age, people had to go to a branch and talk to a bank teller to do their banking business. While the teller took care of whatever issue needed to be taken care of, they learned a lot about the customer through their finances. Based on that information, tellers would offer them products and services they might need. In other words, banking products were sold entirely through personal connections,” Gellért Vinnai, W.UP’s head of product management explains. “With the advent of digitalisation, however, there have been fewer and fewer reasons for people to go to branches so the personal aspect of their relationship with banks have almost completely disappeared.”
So what happened with all that valuable information tellers used to look at to find out what products to offer to customers? It’s still there. Except, without the right data analytics tools, it’s pretty useless.
“Artificial intelligence helps banks rebuild this personal relationship with customers in the digital space using advanced data analysis. With AI tools, they can learn about people as much as possible, the same way bank tellers would do. Or better,” Gellért explains. So, as out-there as it sounds, using machines to personalise human interactions is nothing new. Neither is the technology. Machine learning algorithms are no different to statistical algorithms that have been around for decades. What changed is that we have better access to them, thanks to the rapid development of IT infrastructures and the exponential growth of available information. “Today, we have the means to translate huge amounts of transactional and other customer data into business insights in no time.”
So how to get started with implementing AI and ML within a banking organisation? Gellért swears by a three-step process. “The first step should always be a thorough data cleansing. Get rid of irrelevant or inaccurate records and make sure that what you’re left with is quality data that algorithms can make sense of,” he advises. Once you’re done with preprocessing, you can move on to turning data into actionable business information. “In W.UP’s case, this means data enrichment. We categorise and cluster information and then build customer profiles. Customers are micro-segmented according to their demographic differences but also based on their habits in terms of shopping, travel, sports, entertainment and so on.”
The last step of the implementation process is making these insights work for you. Instead of having to rely on bank employees to segment customers in their head and pick the right offer for them, banks at this point can run algorithms to create micro-segments or even segments-of-one through digital interactions and then laser-target customers with their products. But that’s not all. Gellért adds: “Not only can machine learning algorithms personalise banking offers and digital experiences but they can also help with product development. They allow banks to spot and monitor use patterns among customers and tailor banking products to their very needs.”