FIs use smart automation to improve member experiences behind the scenes, creating more secure and efficient processes while lowering friction. In the “Digital-First Banking Tracker,” Veridian Credit Union’s Brett Engstrom talks about cumulative gains from cutting seconds off transactions, algorithms that learn better ways to spot fraud and automated bots that emulate live assistance.
Automation has become central to modern banking practices, both on the user-facing side and within FIs’ internal operations.
Algorithmic tools such as artificial intelligence (AI), machine learning (ML) and robotic process automation (RPA), coupled with application programming interfaces (APIs), can help FIs make many processes less labor-intensive and remove the risk of human error.
“Today, we’re automating things like transaction cleansing and categorization,” said Brett Engstrom, chief information officer at Veridian Credit Union. “So, basically, within digital banking, it can take every transaction that a member does and say what category that should be placed into so they can better manage [their finances].”
Such systems, with which many FIs and banking customers are already familiar, may also employ ML to learn over time as consumers correct which categories transactions should go into. ML enables smarter automation, permitting the system to recognize not just a specific entity or name, but also patterns and other data that can help it make better categorization decisions over time.
“So it gets smarter over time,” Engstrom said. “And that’s true of a lot of our solutions.”
Learning From the Data
An automated system’s ability to learn and alter its own behaviors is particularly helpful when it comes to reducing risk and improving security, Engstrom said. Such systems can not only make transactions more secure, but they can also spot fraud before it is even committed. They also can help prevent legitimate transactions from being incorrectly flagged, ensuring a better user experience overall.
“Historically, ‘user friendly’ and ‘secure’ have been almost mutually exclusive, with institutions trying to find that sweet spot between ease of use and protecting members and the credit union against fraud,” Engstrom said.
Automated systems with ML can make smarter decisions, going so far as to analyze the unique way a member tends to use an interface to confirm whether a transaction is legitimate. Data such as IP range, time of day and browser or operating system can more positively identify a user, but so can the speed with which a member moves from one click to the next.
“We’re making all these observations about who they are and how they interact with the technology,” Engstrom said. “It’s really the most user-friendly way to protect members: behind the scenes and without adding friction, when they go about their normal interactions.”
That also factors into such situations as card use during travel. Smarter automated systems can see the difference between a suspicious transaction and one that fits a spending pattern typical of a member who is on vacation or traveling for business.
Taking Automation to the Next Level
Smart automation can also help improve the consumer experience even when all it does is take a few seconds out of a manual task. Engstrom noted improvements to Veridian’s call center as an example. With an increasing number of members calling in for services and tasks they may have handled at a physical branch before the pandemic, automation of even small or repetitive tasks can significantly improve the customer experience and free up staff.
“We’re really working to automate more of that with RPA, as far as shaving precious minutes off of every transaction,” Engstrom said. “We’re making sure that we can spend our time where members need us most, instead of doing routine tasks. We can use RPA to take something that used to take 30 seconds and that now only takes maybe nine seconds per transaction.”
Even improvements of a few seconds quickly add up over tens of thousands of calls, Engstrom said.
“We have a lot of automation that’s more like canned responses for the easy answers, the ones people ask 1,000 times every day, so we’re not using a live agent for everything,” he said.
Automated bots that sound and interact the way a live operator would are key in that implementation so that members do not become frustrated or lost in endless phone trees.
Creating Connected Automation
Making systems more efficient goes beyond shaving seconds or even minutes off internal processes. Tying all those systems together also helps eliminate manual processes and reduce fragmentation, and this is where APIs come in.
“Almost all of our solutions at this point are interconnected via API on the back end,” Engstrom said.
Whether those are purely internal systems or connections to partner products, APIs help ensure that customers have a seamless experience. They also allow many services to be addressed within a single user experience, rather than shuttling members back and forth between systems.
The need for intelligent, automated digital systems is not showing signs of waning, Engstrom said. The pandemic pushed consumers toward digital and voice channels, but now that they are familiar with them, consumers are sticking to them for convenience and ease of use. By automating processes in a way that accentuates those experiences, FIs can continue to gain traction with consumers.