As mental health claims surge, Accenture Canada’s Puneet Chattree explains how insurers and plan sponsors are turning to AI to detect early risks in employee health
Mental health-related disability claims have become one of the fastest-growing challenges facing both insurers and plan sponsors in Canada. As the industry grapples with this surge, many are turning to AI as a tool for early detection, smarter claim processing, and scalable support models.
But even as the technology advances, the tension between automation, human oversight, and data security remains front and center. Puneet Chattree, who works with major group benefits providers across North America, sees AI as an essential lever for getting ahead of the growing volume and complexity of claims, particularly in the post-pandemic landscape where employee wellbeing has moved to the frontlines of corporate priorities.
For plan sponsors, the shift is from reactive claim handling to proactive health investment.
“Mental health and chronic disease continue to be two of the fastest growing drivers of disability claims in Canada,” said Chattree, insurance industry lead at Accenture Canada. “It’s actually not just in Canada because we’re seeing it in the US too… Mental health continues to be, outside of chronic disease, a very critical area of focus and an area that requires a lot of early risk signals that can benefit plan sponsors if they if they really think about how to start detecting this and preventing this early on.”
For plan sponsors, Chattree believes that AI gives them a powerful advantage when it comes to identifying workforce risks before they escalate into disability claims. He argues that predictive analytics can help employers spot early warning signs that might otherwise go unnoticed, particularly when they combine structured internal data with less traditional sources.
Rather than just tracking department-level absences or known demographic risks, Chattree believes the real opportunity lies in layering in behavioural data and external signals. For example, economic shifts like rising tax rates or housing costs can exert pressure on certain employee groups. When those macroeconomic stressors are factored into predictive models, employers can better assess who is likely to experience strain and where support might be needed.
He added that geographic disruptions, like wildfires displacing employees, are another overlooked variable that AI can help account for in risk assessments.
He also sees untapped potential for plan sponsors to use data more strategically in managing employee health, notably when it comes to identifying early indicators of disability and mental health challenges. He argues that many employers are sitting on valuable insights that aren't being put to work.
By analyzing internal data, such as departmental absences, usage rates of wellness platforms, and demographic breakdowns, plan sponsors can identify where risks are emerging. Chattree also acknowledged that mental health stressors aren’t always direct as fertility challenges, bereavement, or caregiving responsibilities can all impact employee wellbeing, even if they fall outside of traditional mental health categories.
He believes predictive analytics can help employers pinpoint high-risk groups and tailor support efforts accordingly. If done well, this approach can be strengthened through partnerships with insurers who already run sophisticated risk models on disability trends.
“There’s an opportunity that I don’t think has been fully tapped into, in the way they can start leveraging some of that data that these insurers have,” he said, pointing to a growing shift toward early, targeted intervention.
From expanding access to virtual therapy and coaching, to offering flexible work-from-home arrangements and even subsidized pet insurance, employers are beginning to explore more personalized support strategies.
“There’s a lot of opportunity to start going to what I would call more targeted sort of intervention early on and promoting that in terms of the way that you provide benefits to your employees,” he said, particularly for employees in high-risk segments.
Chattree emphasized that nearly all major group benefits carriers in Canada and even some in the mid-market are actively investing in generative AI, with claims management as their top priority. While he doesn’t name firms directly, he confirms that “all of the top ones are going down this journey right now,” with varying levels of maturity.
Additionally, many of these insurers have already established internal centers of excellence focused on AI development, and Chattree noted that building business cases and demonstrating value is currently a key focus.
“There’s actually a lot of within-year gain,” he said, emphasizing that the benefits of AI adoption aren’t limited to long-term planning.
“It’s not just about the opportunity... it’s how they’re able to use [AI] to really support more on the prevention and the health navigation mandate.”
Chattree emphasized that much of AI’s value in claims management lies in augmentation, enabling smarter, faster decision-making for claims managers, and moving the role from transactional processing to advisory and coaching support.
He explained that AI is now being applied in several stages of the claims process. At the intake stage, it’s helping to synthesize medical documentation, identify relevant coverages, and spot early risk flags that might not be visible to human reviewers. This reduces manual effort and allows for quicker triage.
Still, he cautioned that the foundation for using AI responsibly in insurance begins with strong data governance, something he says should be in place regardless of how advanced the technology becomes.
“We talk a lot about having very strong data control, data governance and a strong cybersecurity framework,” he said. “These things will always stay true regardless of AI or not.”
He noted that companies already disciplined in how they manage and secure data are better positioned to navigate the added complexity that comes with deploying AI. But as the technology evolves - from generative to agentic AI - the pace of change is outstripping existing governance models, particularly when it comes to explainability in claims decisions.
Transparency is also a critical issue for claims, he noted, adding that carriers are becoming more focused on building responsible AI standards, including the ability to back-test algorithms and explain how decisions are made.
Chattree ultimately sees AI as a support system and not a replacement for case managers and agents. He stressed the importance of what he calls the "advocacy role," where professionals act as navigators for individuals moving through the claims journey.
That role must adapt to the preferences of the claimant as some will lean toward a fully digital experience, while others will want a human touch to explain next steps and offer reassurance, he said.
He believes this personalized, concierge-style approach, combining early intervention, access to coaching, and tailored return-to-work plans can’t be automated.
“Humans are always going to have a critical role in human intervention and interaction,” said Chattree. “The need for humans to be in the loop is really, really critical.”
