Inside the Canadian insurer's disciplined wager that the future of disability claims belongs to case managers with better tools, not just algorithms with final say
Disability insurance work is fundamentally about people making careful decisions at moments of disruption in a client’s life. A case manager reviews medical information, considers clinical perspectives, listens to the claimant’s experience, and applies the terms of the contract or policy to determine eligibility and next steps.
It is complex, judgment‑based work carried out in partnership with clients and it is now being reshaped by rapid advance of artificial intelligence.
At iA Financial Group, the transformation is underway, but it does not look like the version sold at tech conferences. There is no autonomous claims engine but rather a deliberate effort to first strip administrative drag out of a profession that has been drowning in it for years, and a working thesis about where machines belong in a process that turns on human judgment.
Cheryl Nicholson, Senior Director, Disability Management, Group Benefits and Retirement Solutions at iA Financial Group, an occupational therapist who has spent three decades in disability management and rehabilitation, is candid about the limits of what AI can do in her field, and equally candid about where it is already paying off.
Everyday work generates tangible gains
The most concrete gain for the organization so far has been practical rather than dramatic. A post-call note-taking tool now handles documentation that case managers used to type while simultaneously trying to listen to claimants. The pilot team’s reaction was pragmatic rather than enthusiastic: they could focus on the conversation, and their file decisions moved faster.
Another initiative, still in exploration, deploys an AI agent to read incoming documents, categorize them, and route them to the right place in the file. That work has historically been done by administrative staff, one document at a time. If it scales, it removes one of the largest fixed costs in claims operations.
Every hour reclaimed from clerical work is an hour redirected to the aspects of disability case management that truly require human involvement: thoughtful conversation, client support, critical analysis, collaborative problem‑solving and professional judgment.
A 50% time savings at the hardest point in a file
A more consequential pilot at the company targets the moment in a long-term disability claim where stakes are highest: the 24-month change-of-definition mark. At this stage, most policies require a comprehensive reassessment of ongoing eligibility, as the criterion for continued benefits shifts and the file must be reassessed in full. Nicholson reports that an AI tool currently being piloted is on track to cut roughly half the time case managers spend on that review.
“When we’re looking at periods like the change of definition, often at the 24-month point in a file, that’s a significant point for disability management,” she says. “And that is looking to save case managers about 50% of their time.”
The mechanism is straightforward. The tool ingests medical documentation, builds a timeline of medical and rehabilitative treatments, identifies which physicians recommended which interventions, and surfaces the file in a form a case manager can absorb in minutes rather than hours. While hallucinations remain a known risk, they are not the norm: the team treats accuracy as a hard KPI and continues to evaluate the tool accordingly.
A 50% reduction in case manager time at the most labor-intensive point in a disability file is a structural change in how a claims operation can be staffed and how quickly claimants and plan administrators get answers.
The distinction that matters: Augmented, not artificial
An articulation of AI that Nicholson finds meaningful is quite simple: this is augmented intelligence, not artificial intelligence. Nicholson returns to it repeatedly when describing how the technology is being deployed in case management.
The distinction is more than semantic. Artificial intelligence implies substitution. Augmented intelligence implies leverage on an existing professional. In a regulated, judgment-heavy line of business such as disability insurance, that difference carries operational and legal weight. It dictates which decisions a model is allowed to influence, which it is allowed to make, and which it is kept away from entirely.
Nicholson is direct on this point. AI does not substitute for a case manager’s decision-making. It is being used to bring information forward in a digestible, fast format, to free practitioners from low-value repetitive work, and to be deployed iteratively with consideration of how each use case fits within the firm’s internal architecture.
The decisions remain with humans, and the firm’s tiered governance approach brings in IT, compliance, and risk partners as use cases involve more sensitive personal information. This ensures long‑term success for group clients and employees while maintaining oversight without displacing judgment.
What the bullish AI narrative overlooks
The default Silicon Valley framing of AI in insurance assumes that case management is a process waiting to be automated end-to-end. Nicholson’s read, from inside the work, is different. “Disability management is very complex. And it’s not a linear process. There’s a lot of nuances,” she says. “It requires a lot of human expertise and judgment. From my perspective, case management is always going to require a human element.”
The human element, in this account, is not a bottleneck to be removed. It is the product. What AI does is allow that human element to scale. A case manager who is not retyping phone calls, not hand-sorting documents, and not re-reading two years of medical history at the change-of-definition mark, is a case manager who can carry more files, respond faster, and spend more attention on the claimants whose situations actually require it.
Nicholson is unsentimental about where this is heading. “I don’t believe that the human elements, the care, the human empathy, the support, the judgment that’s needed by people to manage within the space of disability, are going to change,” she says. “I just feel that it’s going to be incredibly augmented and focused, because we should be able to eliminate a lot of the day-to-day administrative, repetitive, low value work.”
For Nicholson and iA Financial Group, the promise of AI lies not in what it replaces but in what it makes possible: deeper conversations with claimants, greater capacity for thoughtful analysis and sound judgment, and a stronger focus on the supports that truly aid recovery—ultimately improving outcomes for plan members and the organizations that support them.
That is a meaningful efficiency gain. It is not a replacement story.
This article was produced in partnership with iA Financial Group
