Can AI Turn Banks Into FinTech Firms?

There are no incumbent banks that are not paying attention to what the challenger banks are doing. What might have been written off even five years ago as a fad for digital financial services companies, or a niche product for a subset of consumers, is now “top of mind” for banks of all sizes, Jim Priestley, chief revenue officer at Feedzai, told PYMNTS in a recent conversation.

“Even in the last two years, there has been a big shift in legacy banking to [become] tech companies,” Priestley said. “A lot of them have been saying it for a while, but more and more of them are actually building data science and technology teams that really get it.”

What they are getting is that consumers want, and increasingly demand, a suite of services on the front end that make the financial management of their lives easier and more seamless. That includes, at the most basic level, mobile and digital banking. A full 90 percent of consumers said their top online activity was personal banking, according to the latest edition of PYMNTS’ Digital Banking Tracker — and 60 percent reported regularly using at least one financial services app.

However, it extends from there, Priestley noted, to a host of other services around lending, financial planning and deposits. Consumers aren’t looking for one-off solutions for individual financial services; they are looking for an end-to-end journey that is both simple and secure.

For banks to provide those services, it is going to take more than desire and will to get it done. Legacy banks, he explained, need to get free of their legacy systems, as well as embrace artificial intelligence (AI) and machine learning technologies that build a more secure experience on the back end, and a more accessible experience on the front end.

The Legacy Lag

“At the end of the day, rules-based systems are just primitive,” Priestley said, noting that this was the simplest way to sum up the problem with most banking legacy systems. They can only prevent fraud or money laundering on the basis of what they’ve been programmed to flag. They can’t adapt, they can’t recognize new tactics, and when problems are found, they are slow to respond.

It can take months after banks realize they are losing lots of money, or have a damaging consumer experience, to make changes in their legacy systems to compensate. That is not a workable time scale in the modern fraud environment.

“This isn’t a kid in a basement in Minnesota stealing sneakers — these are organized, global criminal organizations capable of launching [a] series of bot armies against every channel probing for weakness,” he said. “The faster your fraud model works is a differentiator.”

What AI and machine learning bring to the table for financial institutions (FIs) — particularly the kind of unsupervised AI that Feedzai favors in its own products — is being able to take the information from the data sets they have been trained on, and incorporate (and iterate off of) new data as it comes in. What that essentially means (from a consumer banking point of view) is that, in real time, the system can make a determination transaction by transaction, instead of being handcuffed to a set of rules. It makes it much harder for fraudsters to disguise bad transactions as good ones because the system is more able to spot emerging patterns in fraud.

The same thing, Priestley noted, can be said of applying AI and machine learning to anti-money laundering (AML) and know your customer (KYC) issues. It’s a massive problem, given the $2 trillion laundered through the banking system every year. It’s not a new problem (banks have always been vulnerable to money launderers), but people are more attuned to it now. Not to mention, due to the rapid proliferation of digital on-ramps into the financial services ecosystem, there are many new opportunities for criminals to sneak their money in for cleaning.

“Things have changed pretty dramatically over the last five years, and a lot of these archaic systems make it easy for criminals to work around them because they aren’t suitable to emerging channels, and [are] unable to handle all of these new data types,” he said. “They just aren’t as mature.”

That is costly for banks in terms of fines, which are steep — and also in terms of the reputational damage that comes with the adverse media that tends to follow money laundering events. Better systems can stop the fraudulent depositing so that the AML violations aren’t happening, and help banks better track the adverse media.

However, the value of adopting better systems based on AI and machine learning, he said, isn’t just about preventing bad things from happening — though that is important. It is just as much about making great things happen.

The Smarter Experience

When Feedzai talks with its partners in banking and digital commerce, the two leading concerns it hears about are security and consumer experience. While they are concerned about security, they have a stronger feeling that it can be managed on the back end with the right technological upgrades and partners. The bigger concern, nearly universally, is consumer experience on the front end.

While these people invariably want to stop fraudsters, they almost equally don’t want to do so in a way that compromises consumers’ happiness. No one wants fraudulent transactions on their account, he said, but consumers also don’t want to be blocked from a purchase they want to make by an overzealous security protocol. That means logging and tracking false-positives so that the system can avoid them is critical.

“But when we look at our banking partners, we see they are taking [this step]. Players like Lloyds and Citibank are thinking about the total experience — not just about log in or starting an account, or lending, but about building a full data set across their channels that can make things work better for their consumers,” Priestley said.

That might mean, for example, the system recognizing a high-net worth, and sending them messages about savings or investment vehicles. That could be noticing new parents, and directing them in how to start a college savings account. It’s not just about managing risk, he said, but about finding how to boost the top line in ways that seem natural to customers. That is a trend emerging in all banks, he said.

It is another bit of evidence to add to the mounting pile that the world of financial services is rapidly evolving into a different type of technologically moderate ecosystem — where legacy and challenger banks have much to learn from each other still. Legacy banks need to develop more of an ability to act like fast-acting, fast-interacting tech companies. Challenger banks can learn how to go from a single service — like, say, student lending — and become a multi-service offering provider. The broad approach is difficult, he noted, and banks know an awful lot about it.

Mostly, Priestley added, the future will be about how new and old financial players are going to collaborate — and find new ways to bring forward better innovations in security and seamlessness.

“As they collaborate together, they can better work to answer how they can share all of this data with each other,” he said, “so that [they] can all fight the bad guys out there together.”

 

To read more about challenging banks, we recommend you to read this: Data will drive the banks of tomorrow

Leave a reply

Please enter your comment!
Please enter your name here