TOFIS Keynote: National Bank of Canada Uses AI to ‘Brew’ a Richer Blend of Banking Services

Gouslisty described his efforts to transform National Bank of Canada, using AI to improve the customer experience, comparing the challenge facing banks around using AI to offering greater value to clients to the coffee industry:  In addition to savings and efficiency benefits, innovation and better customer experience translates to being able to charge higher fees for services, he said. For example, a cup of coffee beans may cost one cent per cup. An individual-serve packet of coffee may cost 25 cents. A cup of coffee at a fast-food chain like Dunkin Donuts may cost $1.50, while a cup of coffee at a “premium” coffee shop chain like Starbucks would cost even more. In short, customers are willing to pay more for higher perceived value.

The increases in data volumes and availability, and increased analytical power means banks can move up the value chain from providing static reports to analysis, forecasts, and ultimately using machine learning and artificial intelligence.

Specifically, Gouslisty outlined four pillars of an AI-based transformation: delayed, structured data moves to being real-time structured and unstructured data; static reports become dynamic performance management assessments; models and forecasts become predictive and prescriptive analytics; and manual work gets augmented or replaced with interactive components, such as robotic process automation and “bots.”

As a result, Gouslisty says the next generation of “talent” won’t just be individual skilled professionals, but will comprise a mix of humans, data, algorithms and intelligent software. And instead of focusing on deriving meaning from data, AI will allow firms to focus on opportunities to extract value and manufacture new revenue from it, he adds.

But such a large shift in approach entails some major shifts in the way banks view and value their data assets. “From an accounting perspective, a server has more value than the data stored on it—so that’s a major shift that needs to happen,” Gouslisty said.

“It’s about changing the productivity frontier,” Gouslisty says. “But when you try to have the AI conversation with management, you hit ‘Henry Ford Syndrome.’ Ford would say, ‘If I’d asked people what they wanted, they would have said they wanted a faster horse.’ And in our space, people tend to ask for more reports or faster reports.”

In addition to changing mindsets, the technology environment that firms currently have in place may not be suitable for a shift to AI-based processes, he said, adding that National Bank of Canada created an internal group called Datahub to run the process of adapting specific models or services to incorporate AI. The group delivered 17 projects in 2017, has 15 underway now, and has almost a hundred other areas that it could apply AI to.