The insurance industry is in the grip of a familiar tension. Carriers are pouring capital into AI-driven automation, promising faster underwriting, smarter claims handling and leaner operations. And yet, as Aon's chief executive Greg Case argued at a recent Semafor conference, the sector risks building a powerful engine with no one qualified to drive it.

"It is inconceivable that a winner in the application of AI isn't going to lead with a world-class people strategy," Case told conference attendees. "Inconceivable."

The warning lands with particular force in an industry navigating a talent crisis that predates the current AI surge - and that AI may be making worse. The context is stark. A Jacobson Group and Aon study found that job openings in finance and insurance fell to their lowest monthly level in a decade by December 2025, dropping from an annual average of 281,000 openings to roughly 138,000 in a single month. At the same time, approximately half of the US insurance workforce is expected to retire by 2038, opening the door to more than 400,000 vacancies in roles that depend on deep institutional knowledge.

The collision of those two forces - a thinning talent pipeline and the automation of entry-level roles - is precisely what Case is pushing the industry to confront.

"This isn't about including talent," he said in his closing remarks at the conference. "This is people-led. Talent leads this story. You can't win here without a world-class talent strategy."

The underwriting problem

Nowhere is this tension sharper than in underwriting. AXA XL's chief operating officer has noted that AI is rapidly absorbing the administrative tasks - submission data entry, loss run ingestion, MVR summaries - that have historically served as the training ground for entry-level underwriters. "Historically, administrative work was how entry-level underwriters learned the fundamentals," the COO said. "Now we have to rethink how we develop talent."

Case made the same point in broader terms. Cutting back on early-career hiring in response to AI-driven efficiency gains, he argued, is exactly backward. "The idea that you would pull back on early careers - I'm dumbfounded by it," he said. "There should be opportunity here for everyone, as opposed to 'my gosh, we need to pull back.' It's exactly the opposite."

That view puts Case at odds with some of the industry's biggest operators. Chubb has announced plans to cut up to 20% of its global workforce over the next three to four years as part of a sweeping digital transformation, targeting underwriting administration, claims and operational support. Allianz, meanwhile, has signalled cuts of 1,500 to 1,800 positions within its travel insurance operations as AI reshapes customer and claims processes.

Case's counterargument is not that efficiency is wrong, but that leaders who stop there are leaving the bigger prize on the table. "If you're about pure efficiency, stand by," he said. "The winners are going to drive benefits from that standpoint, but they're going to drive growth, they're going to drive capability building."

A skills gap inside the AI gap

The industry's AI ambitions are running directly into a shortage of the people needed to execute them. As Insurance Business has reported, many carriers still run on older core systems and have strong actuarial and risk functions but lack the deep AI engineering and data science expertise needed to scale pilot programmes into production. "Running a proof-of-concept is one thing," one AI integration specialist told Insurance Business, "but scaling to production requires specialised skills in areas like MLOps, data engineering, and compliance - and those roles are hard to find."

A 2024 survey by Davies found that 63% of senior managers in the insurance industry believe a lack of digital skills is an issue in their workforce, with 30% describing it as "a very serious issue."

Case is familiar with this dynamic from the inside. Aon, which employs approximately 60,000 people worldwide, placed its first Nvidia chip into a business application in 2009 and had committed roughly $1.3 billion to AI-related programmes by 2023 - what he describes as the largest investment in the professional services industry's history. The lesson from that journey, he said, is that technology spend without parallel investment in people fails to compound.

"How do we take specific steps to actually train, retrain, retool our colleagues so they can be effective in this environment?" he asked. "That's going to be the unlock - an enabled capability we've never seen before, an enabled capability by people."

His formula for Aon is instructive: rather than using AI to shrink headcount, the firm is aiming to make its existing 60,000 employees dramatically more effective. "We don't want 30,000 - we want 60,000 more effective," he said.

Skills-based hiring as the new standard

For the insurance industry specifically, Case's call for a skills-based approach to hiring and development carries practical urgency. As Insurance Business has documented, more than one-quarter of insurance professionals today are over the age of 55, while less than 25% of the industry is under 35. The pipeline of experienced talent is narrowing just as the skill requirements of the remaining roles are expanding.

"We need to do more and better in our industry to attract talent that can be part of all these new things we see coming around technology, including AI," Westfield Specialty's chief claims officer told Insurance Business. "The future belongs to those willing to take strategic risks and invest in the next generation."

Case argues that the answer lies in replacing credential-driven hiring with a continuous, skills-based model - one that identifies what capabilities the business needs today and eighteen months from now, then builds towards those targets through targeted training, apprenticeships and redeployment. Critically, he sees this as an expansion of the talent pool, not a contraction of it.

"We have to evolve ourselves as we think about how we bring in, identify talent, bring in talent, train talent, retain talent," he said. "What a better investment would we make to be successful in an AI world if we can enable it through a better talent strategy."

The hyperscalers, he observed, already understand this. The fiercest competition in AI is not being fought over data centres and compute - it is being fought over people. "Who's fighting over talent? Who's fighting every day? Who's the headlines on talent? It's the hyperscalers." For insurers that have historically struggled to compete on compensation with big tech, the implication is that non-financial factors - career development, skills investment, meaningful work - become the differentiating lever.

The pipeline risk no one is pricing

Perhaps the most pointed implication of Case's argument for the insurance industry concerns the talent pipeline itself. Insurance Business has highlighted that many of the roles most exposed to AI automation have historically served as entry points into the industry - the first jobs that gave younger workers exposure to claims, underwriting and customer service before they moved up. As those roles contract, the pipeline of future experienced professionals risks hollowing out. That is a strategic risk as much as an operational one.

The new generation of roles that remains demands a different profile. AXA XL's COO has articulated the challenge directly: "Underwriters and underwriting assistants will need to understand the rules behind the AI tools they use. They need to be able to interpret outputs from agentic AI, explain results, and make judgment calls that the technology cannot."

That is, in essence, what Case is calling for at the enterprise level: an industry that treats talent development not as a cost to be managed alongside AI investment, but as the strategic condition for AI investment paying off at all.

"Every transformation brings with it the seeds of disruption and the seeds of opportunity," he said. "We are working to drive the opportunity faster than the disruption."

For an industry that has spent the better part of three years debating how much AI will change the work, Case is making the case that the more important question is who will be equipped to do it.

The underwriting problem

A skills gap inside the AI gap

The widening gap: AI investment vs workforce training spend

Skills-based hiring as the new standard

The pipeline risk no one is pricing