Businesses typically plan for anywhere from six months to one year ahead, but with the pandemic upending many aspects of life and the economy, the outlook for even the next month, or even week, can be unclear.
Scotiabank’s analytics and risk experts in Global Risk Management have found a way to leverage machine learning to make short-term predictions and use those insights to tailor their solutions and provide assistance to clients as they navigate through these challenging conditions.
The Bank has developed a cashflow prediction tool — called SOFIA, for Strategic Operating Framework for Insights and Analytics — that uses historical commercial banking data, such as deposits, and trends from the past year combined with machine learning to forecast what clients could expect in the next four weeks.
This rolling average, which is updated in real time, gives the Bank a better sense of which clients are more likely to be impacted by the economic downturn and how to best respond to them. For example, relationship managers can proactively approach those whose cashflow may be under pressure and offer help, such as providing information about customer assistance programs or options for short-term lending.
“It’s really intended to be a conversation starter with our customers,” said John Phillips, Scotiabank’s Director & Head of the Credit Solutions Group. “It’s intended to provide us with insights into accounts which may be trending down so that we can get in front of it and have discussions with our customers that are informed by the data.”
The tool also helps the Bank be more efficient, allowing risk managers to better focus their efforts on the right customers at the right time and can speed up the annual review process. The insights generated by this cashflow prediction tool also help inform the Bank’s planning, such as how much to set aside as provisions.
The tool itself was built about a year and a half ago, long before the emergence of COVID-19. Its original purpose was to digitize and speed up the review process for commercial banking accounts, which had previously been done largely manually and annually, said Yannick Abba, Scotiabank’s Vice-President of the Analytics Centre of Excellence in Global Risk Management. The Bank launched a regional pilot of the cashflow tool in Commercial Banking and Retail Banking early in its fiscal first quarter.
When the pandemic hit, the analytics team realized that given the nature of the information the tool provides, it would be useful during this difficult period.
“What we've been solving for really isn’t a new issue, but it's been highlighted through the pandemic,” said Chris Wise, Director of the National Agricultural Credit Unit at Scotiabank. “It's the need to get a banker out in front of a customer when they need us.
“In the past, we've relied on historical reporting such as financial statements from customers. But now through the pandemic, where it's a more dynamic economic environment, this tool really highlights those accounts that are changing so we can direct our resources to them.”
In turn, rollout of the tool was accelerated. Late in the third quarter, which ended in July, it was scaled up in commercial banking nationally, Abba said, while also expanding into retail banking because of COVID-19. The program was also introduced in Peru in August.
In November, the tool was rolled out across the country.
“We will get improved visibility across the portfolio with the program going national, meaning more customers that we can proactively have conversations with,” said Wise.
Abba said potential next steps include expanding access to the cashflow prediction tool throughout the Latam region, as well as improving the accuracy by leveraging additional information, such as payments.
“We’re going to get better on the back end, because we’re going to inject more data into the predictions,” he said.
This cashflow prediction tool is the latest example of how Scotiabank is using digital technology and analytics to inform decisions and manage risk, said Daniel Moore, Scotiabank’s Chief Risk Officer.
The Bank has been building tools that can be accelerated or deployed in many ways to various parts of the organization, making for faster and better decisions, he added. The work and time invested pre-pandemic in the application of artificial intelligence and analytics, like this tool, is paying off, he said.
“Developing these kinds of tools and analytics had already been on our roadmap, but what has been supercharged by the pandemic is the demand side for those analytics,” he said.
Proactive help is what customers expect of the Bank, Moore said.
“Either as individual or business owner, if your bank comes to you and says your account balance is showing stretched liquidity, we’d like to sit down with you and discuss how we can help you out, that’s a highly different conversation than six months later when the client is having difficulties. An early conversation is good for the Bank and good for the customer. That’s how we should be using data.”