As one of Canada’s Big Five banks, the Bank of Nova Scotia is taking an approach to data, analytics, and AI intended to better understand and serve customers, said Grace Lee, its chief data and analytics officer. Her charter is to advance business growth, customer experience, and operational efficiency through the use of AI, machine learning, and data-driven insights at the bank better known as Scotiabank.
The stakes in customer retention are high: Scotiabank has more than 10 million retail, small business, and commercial customers in Canada, as well as 10 million retail and commercial customers in Latin America, the Caribbean, and Central America. The bank has about 90,000 employees and assets of about $1.2 trillion.
Scotiabank’s two areas of AI application
Over the past couple of years, Scotiabank has engaged in an AI strategy that is very focused on last-mile execution, Lee said. “Where we’ve seen other organizations sometimes fail to capture the benefits of AI and machine learning is that it doesn’t necessarily always result in practical outcomes,” she said. “So, you’ll find that sometimes we call it ‘blue-collar’ AI or analytics, but it’s really around making sure that we see the [AI] models through all the way from inception to [deployment into] production.”
And that means that AI is embedded directly into existing processes and delivering real benefits to stakeholders, such as providing timely advice and personalized offerings for customers, creating some degree of efficiency so employees can better serve customers, or enabling the bank to better predict when its customers might be going through some distress, Lee said. “There is a lot more that we can be doing to actively monitor and really understand the behaviours and therefore the needs and preferences of our customers,” she said.
However, AI is not just helping Scotiabank grow and evolve the customer experience by “knowing better” but by being able to “do better,” Lee said. It also gives the bank the ability “to apply AI to automation, whether it’s in a chatbot or any of the other intelligent automation that we would have across our portfolio,” she said.
When it comes to implementation, it’s important for AI teams to recognize that while AI has traditionally meant artificial intelligence, Scotiabank and other organizations, especially in the banking industry, increasingly refer to it as “augmented intelligence,” Lee said. That’s because of how much it really needs to be embedded into existing processes for it to be of benefit to the bank’s customers and employees.
Grace Lee, Scotiabank chief data and analytics officer
Scotiabank
“There’s very little that we would really want to do that would be entirely automated without some degree of augmentation and oversight by a human,” she said. “So, I think that that’s one really large lesson that we learned early on, when we had tried a little bit more for the artificial and not much for the augmented. We found that the receptivity and the impact it was having, while it’s a very sophisticated model, wasn’t really delivering much for our customers or our employees. So that co-creation is super important.”
AI use cases at Scotiabank
Scotiabank is working on the deployment of natural language processing (NLP) to offer an enhanced customer experience. In the first phase of the project, the bank is building a chatbot to handle basic FAQs, Lee said. It is intended to address “common questions clients might have [about] products and pricing, [for example,] that are being directed to a live agent that can be answered via a user interface guided by AI,” she said. “We want to provide a more conversational experience for our customers so that they’re not waiting for minutes or a long time on the phone to reach an agent when their question or inquiry is relatively simple.”
If the chatbot turns out to be effective, it would not only drive a better customer experience but also let the bank operate more efficiently by enabling its customer service agents or other advisors to work on issues that need to be handled by people.
Scotiabank is using AI to improve the customer experience in several other ways, Lee said.
One is through its Global AI Platform, launched in November 2020. The platform is the infrastructure that lets the bank offer customers faster insights and better advice by using machine learning to anticipate and understand their needs. “We have an on-premises component and we have a cloud component that’s rapidly growing. And that’s where we actually conduct the analytics work and house the data that supports [our] AI solutions,” Lee said.
In January 2021, Scotiabank rolled out another AI effort, the Strategic Operating Framework for Insights and Analytics (SOFIA), an AI tool designed to help the bank better understand which retail and commercial customers will be affected by economic uncertainty and how to serve them by predicting cash flow.
Then Scotiabank launched C.MEE in February 2021. C.MEE uses AI and big data to further improve the customer experience. Using the Global AI Platform, C.MEE analyzes data across all customer touchpoints to identify the most relevant advice it can give to a specific customer, then delivers it through their preferred channels.
By taking signals from the activity of the customers, C.MEE is continually learning and understanding more about their financial behaviour as well as where they are in their lives, thus improving the relevancy of the advice, Lee said.
Across all these projects, “AI drives more efficiency and better insight and information through our employee base and ensuring that, irrespective of how much somebody decides to use an assisted channel or not, they’re getting a much more tailored, personalized, and relevant set of offers or services.”
Organizational structure Is key for AI adoption
One of the key reasons Lee said that Scotiabank’s AI strategy works is because of how the bank is structured organizationally, where the key data and analytics leaders report to a common executive.
The bank also has a dedicated CIO aligned to that function who is responsible for the global data and analytics platform. This person also serves as the bank’s conduit to the other CIOs across the organization so, when the bank needs to integrate AI into various technologies or processes, there is someone who can act as the “interpreter,” Lee said.
This dedicated CIO “would also marry the legacy systems that we might continue to see across the bank with our more modern hybrid infrastructure and more modern capabilities that would come alongside an AI engine or an AI model,” she said. That person also “helps to set those requirements in a way that balances both the old and the new and ensures that we’re making the appropriate trade-offs to get some impact for our customers and for our employees.”
Scotiabank’s three-legged stool of data, analytics, and technology for AI
This three-legged stool of data, analytics, and technology has been critical to the bank’s adoption of AI, Lee said. “It’s less of a capability and more of an operating model question, but it has served us very well in ensuring that we are being practical but also ambitious and [that AI is] being integrated into those technology teams and ensuring that we have the right data pipelines built to make it sustainable,” she said. “We’ve built our [AI] models in a way that respects both of those things. It really is a true partnership across those three groups.”
Because Lee’s team needs such a huge amount of data to build these AI models and AI-based processes, this “handshake” between data and analytics is extremely important to ensure that, when the team has needs from an AI modelling perspective, they are joined at the hip with data partners and aligned on the priorities of what data pipelines need to be built. These teams work together to ensure that the analytics teams across the bank have access to high-quality, well-managed data, Lee said.
“We’ve stumbled a few times in our past because we’ve sought to do AI without that partnership with data,” she said. “From a data-availability perspective, we might be able to gather enough data for us to build the model in the first place. But in terms of sustaining it and being able to use it for ongoing process automation or marketing automation or what have you, that became such a resource-intensive, difficult, error-prone process.”
Scotiabank learned that lesson the hard way: through some early failures. What started as a great idea and something around which Lee’s team felt a model could be built turned out to be untenable from a sustainment and execution perspective. But “in partnering better with data and technology, suddenly analytics models not only become buildable but sustainable,” she said.
