Can AI save Canadian insurance from its talent crisis?

There is a phrase that has circulated through Canadian insurance circles for as long as most practitioners can remember: people don't choose insurance, they fall into it. For decades the industry could afford to treat that as a professional in-joke. It is becoming something closer to a crisis.

Canada's property and casualty insurance industry employs an estimated 297,000 people when direct, indirect and induced employment is counted, according to the Insurance Bureau of Canada's 2024 InsurEconomy Report.

It contributes $38 billion to Canada's nominal GDP and pays more than $12 billion annually in federal and provincial taxes and levies. It is, by any measure, a pillar of the national economy. And it is running short of people.

The demographic pressures are not new, but they are intensifying. The Insurance Institute of Canada sounded the alarm more than fifteen years ago that one quarter of the property and casualty workforce might retire by 2017. The warning proved accurate. The industry adapted, imperfectly, and the clock has kept moving. The Institute's most recent Demographic Analysis of the P&C Insurance Industry in Canada, covering 2022 to 2026, identified a lack of qualified external candidates as a key barrier to recruitment - a finding that lands with particular force at a moment when 8.5% of the current workforce is expected to retire within the next five years, including approximately 15% of senior management.

The broader context is equally stark. A 2020 study by ACORD found that fewer than 4% of millennials would consider working in insurance. More recent Canadian data paints a similar picture: only 4% of students list insurance among their top career choices, with young professionals far more inclined toward technology, finance and other sectors perceived as more dynamic and purpose-driven. Meanwhile, the Canadian insurance industry saw a 30% rise in new job postings in the first quarter of 2024. The market is growing. The pipeline is not.

The scale of what is coming can be illustrated simply. PwC Canada, in its 2024 analysis of workforce trends in Canadian insurance, found that 47% of insurance industry chief executives said a lack of skills in their workforce was hindering their ability to reinvent their businesses.

Estimates from the Government of Canada's occupational outlook for actuaries, mathematicians and statisticians suggest that job openings in those categories will significantly exceed the number of job seekers throughout the 2022 to 2031 period. Actuarial occupations, by some projections, are expected to see employment grow by 22% - a rate far outpacing the 4% average across all occupations. The demand is rising precisely as the supply of experienced practitioners contracts.

Swiss Re, in its 2025 Sonar report on emerging risks, identified insurance workforce and skills shortages as a growing global concern, noting directly that should entry-level positions in claims adjustment and underwriting be automated to reduce costs, this could lead to a lack of people with sufficient job-specific experience in the future - increasing operational errors and raising costs for insurers themselves. That warning has a particular resonance for Canadian carriers, where the concentration of experienced talent in a relatively small professional community means that individual retirements carry outsized institutional consequences.

The problem is not evenly distributed across the country. The IBC's InsurEconomy data shows the P&C industry supports over 72,000 jobs in Ontario and over 32,000 in Alberta, when direct, indirect and induced employment is counted using Statistics Canada's input-output model. Quebec's P&C sector contributes $4 billion to that province's GDP. The talent pressures in these provincial markets differ in character: Toronto's concentration of head offices creates a competitive hiring environment that smaller regional players cannot easily match on salary, while brokerages outside the major centres face the additional challenge of geographic thinness in the professional talent pool.

What makes the demographic shift in insurance qualitatively different from staffing pressures in most other industries is the nature of what retires along with the people. Underwriting expertise, particularly in commercial, specialty and complex personal lines, is not a credential. It accumulates over careers, embedded in judgment about risk that comes from years of loss experience, in broker relationships built over decades, in the ability to read a risk presentation and know intuitively what the data does not show.

Research suggests it takes seven to ten years to develop comparable expertise to an experienced underwriter or senior claims professional. The industry faces the prospect of losing the equivalent of tens of millions of years of combined professional experience over the coming decade - knowledge that cannot be replaced quickly or at scale, and that no training manual fully captures.

The brain drain concern is particularly acute in actuarial and underwriting functions, where the Government of Canada's own projections foresee demand substantially outpacing supply for the foreseeable future. But it extends across the industry's operational backbone: claims management, risk engineering, compliance, and the broker relationships that underpin distribution. Canada's regulatory structure, with its provincial patchwork of requirements, adds a further layer of institutional knowledge that departing professionals take with them - and that new entrants must accumulate slowly.

The talent problem is compounded by a skills mismatch that is simultaneously a cause and an effect of the recruitment challenge. A 2024 McKinsey report found that 60% of insurance roles now require skills in data analytics, artificial intelligence or cybersecurity. PwC Canada's analysis of the sector noted that the talent gap is particularly wide in precisely these specialised areas. Yet the graduates and technology professionals most fluent in data science, machine learning and cybersecurity are those most actively courted by Canada's growing technology sector - and least likely to identify insurance as a natural career destination.

Randstad Canada's 2024 research on the finance and insurance sector found that only around 34% of employees felt their current roles provided sufficient opportunity to develop, and that approximately 45% had had no discussions with their employer about reskilling. At a moment when the skills required to practise insurance are shifting more rapidly than at any point in the industry's history, those figures suggest a sector that is not yet investing in its existing workforce at the pace the transformation demands.

The Insurance Institute of Canada's 2023 industry report found that 64% of the P&C workforce is expected to be operating in a hybrid environment within two years - a shift that brings recruitment advantages but also complicates the mentorship and knowledge transfer that has historically been how the industry trains its next generation. Senior underwriters and claims professionals do not easily pass on their judgment through a video call. The informal learning that happens in shared office space - the overheard conversation, the question asked after a meeting, the case review conducted over lunch - is harder to replicate at a distance.

Into this environment comes a substantial new piece of research that, while conducted primarily with American data, bears directly on the decisions facing Canadian insurance executives and human resources leaders.

A survey of academic economists, artificial intelligence professionals, policy researchers and expert forecasters - conducted by researchers affiliated with the Federal Reserve Bank of Chicago, Yale School of Management, Stanford University, the University of Pennsylvania and the Forecasting Research Institute - set out to measure what informed observers genuinely believe artificial intelligence will do to the economy over the coming decades.

The study's conclusions are more measured than the predictions sometimes emanating from technology company chief executives, and for that reason potentially more useful for strategic planning. The median economist surveyed expected annual GDP growth of 2.5% by 2030, only modestly above current baselines - despite assigning a 61% probability to significant AI progress by that date. The gap between those two numbers reflects a consistent lesson from economic history: transformative technologies routinely take longer than expected to show up in aggregate productivity statistics.

The economists' written rationales for that gap are familiar to anyone who has watched technology arrive in financial services. Multi-decade lags between capability and adoption. Workflow redesign. Regulatory adaptation. The time required to build complementary systems and develop the human capacity to operate them effectively. Electrification, the personal computer, the internet - each went through a comparable period in which the technology existed long before its economic effects were measurable. The researchers assigned only a 14% probability to the most dramatic scenario they tested, in which AI systems broadly surpass human performance on most cognitive tasks by 2030. But they assigned a combined 61% probability to moderate or rapid progress of some kind.

For Canadian insurers, that finding has a direct operational implication: artificial intelligence is unlikely to rescue the industry from its talent crisis in the near term. The tools are arriving, but the transition will take years and will require experienced human oversight throughout.

The study's findings on labour markets deserve close reading from anyone responsible for workforce planning in a Canadian insurance context.

Across multiple AI progress scenarios, the economists surveyed ranked 43 occupation categories by expected employment change through 2030. The results at the bottom of the ranking - the roles most consistently expected to shrink - map with uncomfortable precision onto the positions that form the entry-level pipeline for much of the Canadian insurance industry.

General and keyboard clerks ranked last. Customer service clerks ranked third from the bottom. Numerical and data recording clerks - a category that encompasses much of the routine work in policy administration and claims - ranked sixth from the bottom. These findings held even before conditioning on any particular AI scenario. In other words, economists' baseline expectations already place these roles in the job-loss column.

At the top of the rankings sat occupations defined by human judgment, trust, physical presence and contextual reasoning: personal care workers, health professionals, protective services workers. The common thread is the kind of work that cannot easily be reduced to a set of rules and executed by a machine - work that requires reading a situation, exercising discretion and being physically or relationally present.

This distinction maps onto Canadian insurance in a way that is both reassuring and complicated. The roles at greatest short-term risk from automation - routine claims processing, standard policy administration, data entry - are exactly the roles that have been hardest to fill and that have traditionally served as the entry point into the profession. The roles most likely to endure - senior underwriting, complex claims management, broker relationships, risk engineering - are precisely those from which the most experienced practitioners are departing.

A technology that automates the routine tier of insurance work while freeing experienced professionals to focus on judgment-intensive decisions could, in theory, dramatically extend the productive capacity of a shrinking senior workforce. The arithmetic is attractive for carriers struggling to cover workloads with reduced headcount. But the same technology raises a longer-term question that Swiss Re's Sonar report identified directly: if the entry-level positions that have historically trained the next generation of experienced professionals are automated away, where do the senior underwriters and complex claims professionals of 2035 come from? The pipeline problem that automation appears to solve in the short term may be the one it most deepens over a longer horizon.

_{Framework: Statistics Canada (2023). Occupation placements are illustrative, based on published AI exposure and complementarity analysis.}

The most instructive number in the entire body of data on insurance recruitment in Canada is not a workforce statistic. It is that 4% figure - the share of students who list insurance among their top career choices. It sits alongside the finding that the Canadian industry saw a 30% rise in job postings in the first quarter of 2024 alone. The market is expanding. The interest is not following.

The industry is not unaware of this problem. The Insurance Institute of Canada has run career development and awareness initiatives for years. Provincial institutes and chapters have invested in outreach to schools and universities. Carriers and brokerages have developed apprenticeship programmes and graduate schemes. Markel Canada, in a 2024 analysis of generational change in the industry, called for authentic evolution in how the sector presents itself - noting that Gen Z's expectations for purpose-driven work, technological engagement and career transparency are not unreasonable demands but rather a reasonable description of what insurance, properly explained, actually offers.

The difficulty is that proper explanation rarely reaches the people who need to hear it at the moment they are making career decisions. Canadian students, when they encounter insurance at all, typically do so as consumers - buying auto coverage for a first car, navigating a tenant policy, dealing with a claim for a damaged phone. That is not a context that reveals the intellectual depth of underwriting, the complexity of commercial risk management, the scope of the actuarial discipline, or the genuine economic and social importance of the industry to Canadian communities recovering from flood, fire or financial loss.

The economists' study found that experts are most uncertain precisely when the stakes are highest. Under the rapid AI progress scenario - the one they assign only a 14% probability - the range of economic outcomes widens substantially. Labour force participation in the United States is projected to fall from its current level to 55% by 2050, with roughly 10 million of those departures attributable to artificial intelligence. Wealth concentration increases. The share of income going to labour falls. White-collar employment stagnates as a proportion of the overall workforce.

_{Sources: U.S. Bureau of Labor Statistics (historical); Karger et al., "Forecasting the Economic Effects of AI," March 2026. Forecast values from Table 25 and 26: slow 59.2%, moderate 57.0%, rapid median 55.0%, rapid 10th pct. 45.5%, rapid 90th pct. 61.0% by 2050. 2030 rapid band: 10th pct 54.7%, 90th pct 62.3%. Current rate: 62.6% (2025).}

For Canadian insurance, those macro-economic shifts matter beyond the immediate staffing question. An economy in which a growing share of the population is outside traditional employment, in which wealth is increasingly concentrated, and in which the nature of commercial activity is being restructured by automation, is not simply a different labour market. It is a different risk landscape - with different exposures in commercial lines, different personal lines profiles, different demands on group benefits and disability coverage, different questions for the actuarial models that underpin pricing across the industry.

The P&C industry is already navigating one structural transformation - the accelerating impact of climate-related natural catastrophes on claims frequency and severity. Canada's insured losses from severe weather totalled more than $3.1 billion in 2023, making it the fourth-worst year on record; 2024 then shattered all previous records with $8.5 billion in insured losses, nearly triple the 2023 figure and more than twelve times the annual average recorded in the decade between 2001 and 2010. The industry that will need to price, underwrite and manage those risks through the 2030s and beyond is the same industry that is currently struggling to attract the people who will be doing that work.

Canada's general insurance market is projected to grow from approximately $87 billion in gross written premiums in 2024 to over $115 billion by 2029, according to GlobalData forecasts. That is a market expanding at a compound annual rate of 6% - a trajectory that will require more people, more expertise and more institutional knowledge, not less, over the coming decade.

The industry has the benefit of knowing the challenge ahead of time. The Insurance Institute of Canada's demographic research has been documenting it since 2007. The numbers have not dramatically improved in the intervening years, but the urgency has. The retirement wave that was a projection in 2007 is an active reality in 2025. The skills gap that was identified as an emerging concern is now a recognised constraint on carriers' ability to grow.

Artificial intelligence may yet reshape parts of this picture. The economists' study documents genuine uncertainty about the speed and scope of that reshaping - and identifies the automation of entry-level positions as a specific risk to the future talent pipeline, not only an opportunity to reduce near-term staffing pressure. The industry's task is to hold both of those realities simultaneously: to deploy the tools that are becoming available without dismantling the human infrastructure that makes experienced judgment possible in the first place.

The phrase people fall into insurance has survived for so long partly because, once people do arrive, they tend to stay. Retention in the industry has historically been strong; the pipeline problem is far more one of initial attraction than of subsequent commitment. That asymmetry suggests the solution, while not simple, is at least identifiable. The industry needs to find people before they decide it is not for them - not after.

That means being present in universities and colleges with something more than a booth at a career fair. It means making the case, clearly and confidently, that analysing catastrophe risk, pricing cyber exposure, managing a complex commercial claim or building actuarial models for an industry serving millions of Canadians is exactly the kind of technically demanding, socially purposeful and intellectually substantive work that ambitious young professionals are looking for. It means, in short, telling the truth about what insurance actually is - and doing so before the competition does.

The clock, as the demographic projections have been saying for fifteen years, is running.