.jpg)
Social Media
Follow us on our social media channels
Contact Us
Thank you! We will get back to you as soon as possible.
Oops! Something went wrong while submitting the form.
I lead the advanced analytics exploration team at Wawanesa. Wawanesa is a Canadian mutual insurance company, one of the larger property and casualty insurers in the country, owned by our policyholders rather than shareholders, and named after the small Manitoba town where the company started more than a century ago. My team's job is to find where advanced analytics and AI can genuinely help the business and then build it. Which makes it a little awkward to admit how long it took me to use these tools on my own work.
If I'm honest, the value was never hard to see in the obvious places. AI as a coding assistant for myteam, AI as the solution sitting inside a business process, AI for the routine things like summarizing a long thread, taking minutes, or getting a first draftof an email started. What I couldn't quite picture was its place in the work that actually needed a director. The judgment, the strategy, the messy humanparts of the job. That, I assumed, was mine alone.
Somewhere on the way to a director title, I'd also filed my technical days under "past life." I came up through the build, but I wasn't the one in it anymore, and I'd made my peace with that trade. So what surprised me most about actually using these tools is that they pulled some of that technical muscle straight back out of retirement, just not in the way I expected.
Because getting something genuinely useful out of a model is not "type a question, take the answer." It's a craft, and a technical one. How you frame the problem.What context you decide is worth giving it, and what you leave out. How you break a tangled question into parts a model can actually reason about. When to accept the answer, and when to push back and say that's the easy version, try again. The prompt is the easy part. The thinking behind the prompt is the work.I had to relearn it, and I was rustier than I'd like to admit. For the first little while my prompts were careless, and my results were exactly as good as Ideserved.
What changed was treating the model less like a search box and more like a problem to set up well. And that brought me to the part I genuinely didn't expect. The place that skill pays off most is not code. I have a talented team for the building, and I'm not trying to take that work back. Where AI has earned a permanent place in my week is somewhere else entirely, as a challenge partner for ideas, concepts, and plans.
Much of my team's work starts as an exploration, a short, scoped project to test whether an idea is worth building before we commit real time to it. Before I take one of those forward, or take a idea to my own leadership, I'll lay it out and ask the model to pull it apart. Argue the other side. Find the assumption I've leaned on too hard. Tell me where this plan breaks first. Recently I was set on green lighting one exploration. I handed the model my full case for it and set it loose on the other side, and the rebuttal was sharp enough that I reordered part of my roadmap before anyone else had even seen it. That is not the model handing mean answer. It's the model making me defend my reasoning while it's still cheap to be wrong.
And the quality of that exchange tracks exactly with the technical skill. A lazy prompt gets a flattering, useless answer, because agreeing with me is the path of least resistance. A carefully built one, full of real context and real pressure, gives me an actual sparring match. The relearned craft and the strategic value turned out to be the same thing wearing different clothes.
I assumed leadership meant leaving the technical part of the job behind. It just changed shape. I'm not writing code all day any more. I am pulling a plan apart before it's expensive to be wrong, and finding the technical part of my brain very much still required, only now pointed at ideas instead of syntax.
If you work in data and analytics, in Calgary or anywhere, that might be the most useful reframe I can offer. These tools aren't only for the people still writing the code. Used well, and using them well is a real and learnable skill, they make the rest of us sharper at the thinking that was supposed to be our job all along.