The conversation around AI and insurance jobs in Canada has shifted, and according to Sonia Sood (pictured), national human capital insurance leader at Deloitte, the industry has moved past a simpler question of whether to deploy AI toward a more fundamental one about how work itself should be reimagined.
"Two years ago leaders were asking how might I deploy AI into my organization, and now I'm seeing more of my clients actually asking the question, how can I actually reimagine the work that my people are doing, so that it's not just AI for the sake of it, but it's actually how do we have our people do work better with AI?" Sood said.
That distinction sits at the centre of how she advises clients across both life and health and property and casualty insurers. Rather than treating AI as a tool bolted onto existing roles, Sood said the more productive approach starts by breaking a job down into its component parts.
"I've been guiding my clients, from CEOs to C-suite leaders, on how we can use this as an opportunity to take a job, whether it's an underwriter role or a claims adjuster role, and ask what all the activities are that this role does today," she said. "What are the things that could simply be automated? Where are there practical applications for AI? And what is the work we actually don't want our people doing, because we want to reposition them toward work where human skills matter more?"
Sood said usage of AI is already widespread across the insurance value chain, in claims, customer service, underwriting, corporate functions, and back-office operations. But she pushed back on the idea that this amounts to a wave of job elimination.
"For me, the hype right now is around AI making large-scale job elimination," she said. Without organizations first taking the time to reimagine how work actually gets done, Sood said widespread job elimination is unlikely to happen on its own, unless a company was already headed that way through natural attrition or had already identified high-cost areas it was planning to scale back regardless of AI.
Where organizations get this right, in her view, comes down to deliberateness. She said the goal should be augmenting workers with AI while protecting the parts of the job that depend on distinctly human capabilities.
"Things like human judgment, decision making, empathy, those are capabilities that are going to be very difficult to replace with AI," Sood said. "We're going to always need the human aspect of that."
She pointed to claims specifically as an example of where that balance matters most. Deloitte's research, she said, shows that customers still want a mix of digital and human touchpoints, particularly during moments that carry emotional weight.
"Customers actually want an omnichannel experience, particularly for the moments that matter when they're going through a claim that is deeply personal and emotional," Sood said. "They need that support and they need that empathy to come from a human who understands what they're going through."
What organizations choose to do with the efficiencies AI generates, she said, varies considerably and isn't something she'd generalize across the industry. Some reinvest those gains into new capabilities or further technology. Others, facing compressed margins, use them to reduce their workforce. Sood said the underlying business factors driving those decisions are highly case-specific, shaped by an organization's starting point, its financial performance, and its strategic priorities.
She was careful to separate that broader industry pattern from any one company's specific circumstances, noting she couldn't comment on the reasoning behind any individual organization's workforce decisions. But she said the general logic follows a predictable path when a company's operations are concentrated in the same areas where AI use cases are most mature.
Sood said Canada's insurance sector, as a whole, is likely to be more cautious than some international peers when it comes to large scale workforce reductions tied to AI, at least in the near term.
"I would see Canada being much more conservative in any sort of wide-scale workforce reductions," she said, pointing to Canada's regulatory environment and labour conditions as contributing factors, alongside a slower pace of digital and AI adoption compared to European markets.
That caution, she said, doesn't mean Canadian insurers are moving slowly on AI adoption itself. She said she's seeing increased investment in generative AI, along with growing use of automation in claims, underwriting, customer service, and technology delivery. What concerns her more is whether insurers are doing enough around the surrounding infrastructure needed to support that shift responsibly, from workforce reskilling to stronger AI governance.
Even so, Sood said the framing of AI as a threat to jobs overlooks what she sees as the more accurate story unfolding inside insurance organizations today.
“The story here in the near term is more about workforce evolution rather than workforce replacement, which I think is where a lot of the hype is coming from," she said. "The idea of large-scale replacement is a bit of a myth. The real question is how we reimagine the work people are doing, so that what we do today looks different a year from now, two years from now."
That evolution, she added, applies as much to leadership as it does to frontline roles, and it isn't a one-time shift.
"This is going to be an ongoing transformation," Sood said. "The world of AI is moving so quickly that what's true today may not even be true by this time next year."
