Scotiabank’s CTO Michael Zerbs reports that the bank currently has several dozen AI use cases. Use cases typically bring in benefits from millions to tens of millions dollars per year. But if these numbers sound impressive, Michael says to keep in mind that they are actually very, very small in comparison to the bank’s $923B in assets: “Yes, AI is a cost revenue opportunity and, yes, every dollar counts. But for now, our AI use cases are about enhancing customer experience and optimizing internal processes one small step at a time. At this point, for us, it’s also really about learning and developing our talent to become more fluent in the language of AI.”
“Think of this technology like the internet back in 1995,” Michael says. “At that time, was it clear that you have to pay attention? Not necessarily, but looking back, it is clear that you should have. AI holds future promise. It’s just a matter of time before this technology fundamentally shifts our strategy. But for now, while it’s still early days, it is essential to engage to be ready.”
Using AI to make localized, operational improvements
At Scotiabank, they are using AI to make highly localized improvements across a range of functions and activities, including customer insight, fraud prevention, collections, and what is known internally as smart observation, or process optimization. “Each case is 1 step out of a long chain of steps,” says Michael.
Take collections. “We use AI to look at historical patterns around customer payment history, such as prior willingness to pay, and predict the most effective course of action for our collections team. Should they call, text, or email a particular client about an outstanding credit card bill? If they should call, what would be the best timeframe for that client? Should the team wait a day, a week, two weeks?” Michael explains the value here in terms of customer satisfaction as well as organizational capacity. “When a customer is a little bit late but would eventually get around to paying all on their own, a reminder call has 2 negative consequences: 1) We annoy the customer. 2) It uses up valuable capacity from within the bank. “We have a finite number of resources to deploy against a very large number of possibilities and challenges,” says Michael. “AI can help us figure out which are the right ones to invest those resources towards.”
AI: the challenge of tomorrow
Michael explains that they are focused on raising the bar for leaders to understand AI and analytics across the bank. “When we discuss AI, we often jump to talking about data scientists and statistics, but it’s also important to talk about how to engage senior business leaders early, so that AI initiatives become business-led rather than technology-led,” he says.