From Hours to Minutes

As part of The Actuary Canada’s ongoing focus on AI and technology, I connected with Mike Wang, FCIA, Director of Shared Actuarial Operations at Sun Life in Toronto. Our discussion traced a series of technological transformations and reinforced the idea that companies may benefit from adaptable systems and close collaboration with technical experts.

I’ve always been curious about your career journey. How did you get started?

Mike Wang: I’ve been interested in technology for as long as I can remember. When I graduated from university (University of Waterloo) and started my first actuarial job in 2008, I quickly noticed that a lot of my tech background could be applied to actuarial work to improve efficiency.

What can you tell us about your experiences applying technology to actuarial work?

Mike: Because of my background, I naturally think about how to make processes more efficient, especially around programming and systems design. Give me an Excel process that handles a moderate amount of data or a model built in Moody’s AXIS actuarial system, and I can usually make it run 80-90% faster.

I’ve worked with so many languages, C#, Java, Visual Basic, SQL databases, that they all start to feel similar. I’ve been building data‑driven applications since the early 2000s.

What can you tell us about advanced data infrastructure development?

Mike: When I started in 2008, in my experience, the data environment was pretty immature. That gave me room to build reporting systems using best practices from a data architecture perspective.

In my view, there are two schools of thought in data modeling:

1. Build the perfect data model, or
2. Throw everything together and structure later (Data Vault methodology)

I started my career in the “perfect model” camp. Later, I realized that the Data Vault was far more effective for me. It lets me build applications quickly without getting stuck designing a rigid data model upfront. This was especially helpful because business needs often change early in a project, causing rework and delays.

What was one of your early major projects?

Mike: When I was with Foresters Financial as a senior actuarial analyst in 2015, I designed a full financial planning and analysis (FP&A) reporting system. It produced the income statement, balance sheet, capital projections and source of earnings (SOE) explanations.

The workflow started with an AXIS run, then the system consolidated AXIS results with manual inputs in a SQL database. AXIS took about an hour; the reporting system took 10 minutes to generate everything.

It worked across all product lines, participating life insurance (par), whole life, term life, universal life and annuities, and gave us great insights. I also handled a lot of their AXIS infrastructure.

What can you tell us about system transformation at Manulife?

Mike: I joined Manulife as an associate actuary with its Group Finance Actuarial Systems team in 2018. I’d heard about their system transformation and wanted to see it firsthand. I joined to help rebuild their Life Insurance Capital Adequacy Test (LICAT) calculation system to meet accuracy, reliability and performance requirements.

I wasn’t expecting to rebuild a system that had just gone live earlier that year, but that’s what happened. Using my experience with data‑driven applications, I redesigned the LICAT system so it could handle 1,000 times more calculations. It was a small team, and I led the effort.

What can you tell us about your experiences at Sun Life?

Mike: When I joined Sun Life as a director, LICAT Initiatives in 2019, the company was using a LICAT system similar to the one I had replaced at Manulife. My first priority was to rebuild it. The new LICAT system reduced refresh time from six hours to 10 minutes.

LICAT depended on a separate data submission and validation system. During the IFRS 17 transition, I rebuilt that too, reducing submission time from four hours to three minutes.

With the new system, we added more new data sources in one month than in the previous three years, while eliminating a significant annual run-rate cost associated with the system it replaced.

It was flexible enough that we implemented the IFRS 17 LICAT guideline changes in under five weeks and reduced support requirements from eight full-time equivalents (FTEs) to two.

A project we worked closely on at Sun Life was the transformation of its Drivers of Earnings (DOE) system, moving from an Excel-based system to one that provided greater granularity, better insights and faster turnaround. What can you share about that work?

Mike: The DOE tool was a big challenge. The old Excel‑based tool could only go down to the business unit level for most of the businesses. During quarter‑end, it ran once every three hours and could take up to an hour and a half. It wasn’t reliable all the time and required around-the-clock manual monitoring.

So, we designed a new DOE infrastructure together. You and your team provided the vision—what functionality you wanted, what future state you imagined—and I designed how to achieve it, did all the coding, and added my own thoughts on how to make it sustainable.

The new DOE system:

• Allowed granularity down to major product and portfolio levels
• Reduced runtime to 3 minutes
• Produced reliable results
• Offered far more flexibility

Was the goal to build in future‑proofing, so that when restatements or upstream changes came a few quarters later, everything was already set up?

Mike: That was the goal: to build something long‑lasting that wouldn’t need constant updates unless new functionality was desired.

What excites you most about AI?

Mike: What excites me most about AI is what happens when something that used to be scarce suddenly becomes abundant.

In my area of work, writing code used to be a major constraint. Large language models (LLMs) can now produce code at a speed and scale that simply wasn’t possible before, opening up possibilities that were previously beyond the practical limits of human effort. And I don’t think this is limited to coding. Across actuarial and knowledge work more broadly, when something that once required substantial human effort becomes inexpensive and available at enormous scale, it does more than improve efficiency—it changes what is possible.

That is what excites me most. We are still looking at AI largely through the lens of the work we do today, but I think the range of what becomes possible will extend far beyond simply doing today’s work faster.

In closing, where do you see future transformations heading?

Mike: AI will drive the next wave of transformation. The capabilities of AI models and the tools they can interface with are advancing incredibly fast. I can imagine a team of AI agents doing what I do, but with far greater efficiency and accuracy.

My goal is to train AI to apply the knowledge and techniques I’ve used throughout my career, but at a much larger scale and with greater impact.

Statements of fact and opinions expressed herein are those of the individual authors and are not necessarily those of the Society of Actuaries or the respective authors’ employers.

Copyright © 2026 by the Society of Actuaries, Chicago, Illinois.