By Michael Hainsworth
Canada’s biggest life insurance company has seen a lot of change in its history. But for Manulife, the biggest change since its founding in 1887 is happening today behind the scenes. Calculating risk has always been a risky business. But after migrating 5,000 servers around the world and 300 risk calculation algorithms into a single cloud-based system, the power of big data is being harnessed using artificial intelligence. Not since the invention of the adding machine a year after Manulife first opened its doors has the company had a more powerful tool at its disposal. And this tool predicts the future.
“We’ve been predicting the future for many decades. I would say we’ve been doing it using traditional analytic techniques,” says Jodie Wallis, Global Analytics Officer at Manulife. “I think the question is, where can we get that edge on accuracy? Where can we get that edge on precision? And that’s kind of the goal when using machine learning, right? We’ve made good predictions in the past. We’ve operated effectively. The question is where can we get that edge and make better predictions, more accurate predictions and therefore make better decisions about how to help our customers live better lives.”
The move to amalgamate thousands of servers and hundreds of algorithms into a single platform helped Manulife cut costs 40% over 3 years.
Microsoft Canada’s General Manager for Azure, Henrik Gütle, says it all starts with knowing where the data is and how to manage it. “Data lives in many places,” he points out. “It’s often siloed on different servers across the organization. Some data might already be on the cloud. Some data resides on people’s machines. And with 134 years of data, some data will be analog in the archives. I think first step is really to get a good sense of what data is available and get the right governance in place.”
Once the data was organized, cleaned, and audited, Wallis found an unexpected benefit of migrating to the cloud. “We didn’t realize the impact it would have on the way we collaborate,” she enthuses. “One of the best unexpected benefits is the collaboration is so much easier. We have people working in over 12 markets around the world. We have people in different departments. We have business partners. We have technology partners. We have data partners and the ability to collaborate has really been one of those great surprises.”
While improved collaboration turned out to be a side-benefit of the migration to the cloud, Gütle says the real strength is in AI’s ability to look at “what-if” scenarios. He says it’s intensive in two areas: computing and staffing. “Which also means you need a high-level set of skills and expertise to actually run these analyses. You’re basically limiting the value of those analyses to a few people inside the organization who actually have the skills and capabilities to do it.”
But will that specialized skill-set may become generalized over the next decade? Wallis suggests the dedicated role of “Data Scientist” will be as anachronistic in 10 years as someone putting “touch typing” on their resume today. “Who comes out of high school without knowing how to type? Who will come out of high school without knowing how to build algorithms? I mean, it’s a great future.”
