Written by Sabine VanderLinden 

Early career job postings in the most AI-exposed occupations have flatlined. In the PwC AI Jobs Barometer data published by the World Economic Forum this June, they are the only quartile where that has happened. And when the Forum ranked 18 sectors by AI exposure at entry level, financial and insurance activities came first. Not technology. Not media. Insurance.

This is the third edition in a series on workforce transformation in the agentic enterprise. The first mapped the ten skills we cannot afford to delegate. The second mapped the ten roles where that ownership will live. This one asks the question sitting underneath both: 

If AI agents absorb the work juniors learned on, where does the next generation of judgement come from?

What Is Workforce Transformation in an Agentic Enterprise?

Workforce transformation in an agentic enterprise is the deliberate redesign of jobs, skills, career pathways, and performance measures so humans and AI agents create value together. It determines which tasks agents absorb, which responsibilities humans retain, and how the organization rebuilds its capability pipeline as digital labor takes over routine execution.

That definition carries a warning. Most insurers are doing the first half, absorbing tasks, at industrial speed. Almost nobody is doing the second half, rebuilding the pipeline, with the same intent.

The First Rung of the Agentic Frontier Is Disappearing

Globally, 37% of young workers are employed in occupations with medium to high exposure to AI-driven task change: 3 in 4 across Eastern Asia, 2 in 3 across Northern America and Europe.  

Insurance sits at the top of the exposure table because its entry-level work is exactly what agents do well: routine, information-rich, high-volume processing.

The hiring data confirms the squeeze. Across advanced economies, early career postings in high-exposure occupations have flatlined while every other quartile recovers. In financial services specifically, more than half of CEOs expect AI adoption to reduce junior headcount within three years. For senior roles, that figure is roughly 1 in 10. Expectations of AI-driven structural change at entry level run almost twice as high as for mid or senior levels. 

Rigour matters here, because rigour is part of the human advantage. 

Those posting declines began nearly a year before ChatGPT launched, and leaders interviewed by the Forum admit AI often serves as the visible explanation while economic uncertainty and cost discipline do most of the driving.

So this is not yet a story of agents deleting jobs. It is a story of insurers quietly choosing not to rebuild the bottom of the pyramid while they automate it. Not a hiring slowdown. A structural choice, made one headcount freeze at a time, often without anyone deciding it.

The agentic frontier is not removing your entry-level jobs. Your planning cycle is, one unfilled requisition at a time.

Digital Labor Is Eating the Apprenticeship

Ask your best underwriters how they became your best underwriters. The answers are strikingly consistent: triaging broker submissions, working the referral queue, handling the small claims nobody fought over. Repetition built pattern recognition. Pattern recognition became instinct. Instinct, tested against consequences, became judgement.

Those are precisely the tasks digital labor now absorbs. The traditional learning-by-doing model is weakening, and the Forum's interviews surface the structural worry beneath it: capability can no longer be assumed to emerge organically through job hierarchy alone. One executive put the market failure plainly: 

"If we all stop building entry-level talent, we won't be able to buy it elsewhere in the future."

The macro numbers sharpen the stakes. The Future of Jobs Report expects 39% of workers' skills to be transformed or outdated by 2030, and 63% of employers already name skill gaps as the biggest barrier to transformation. Of every 100 workers, 59 will need training by 2030, and 11 are unlikely to receive it. In insurance, those 11 are tomorrow's missing underwriters, claims leaders, and portfolio stewards.

Here is the uncomfortable parallel. An insurer that automates its first rung without redesigning it is managing its talent book the way a withdrawing carrier manages a closed book. No new business, falling expenses, improving ratios. Run-off looks healthy every single quarter, right up until the book runs out.

Entry-level workers can feel it. Around 28% believe half or fewer of their current skills will still matter in three years, and entry-level roles in the highest AI-exposure quartile show nearly twice the rate of net skills change of other roles. The pipeline knows it is being rebuilt. The question is whether anyone is holding the blueprint. This is workforce transformation happening by default when it should be happening by design.

Four Futures for Digital Labor, One Axis You Control

In January 2026, the Forum's scenario team published Four Futures for Jobs in the New Economy, mapping 2030 along two vectors: the pace of AI advancement and workforce readiness. The four futures that result (Supercharged Progress, the Age of Displacement, the Co-Pilot Economy and Stalled Progress) read very differently, and executives are already split on which is coming: 54% expect AI to displace existing jobs, 24% expect it to create new ones, and just 12% expect it to lift wages.

Look at the axes, though, and the strategy question simplifies. No insurer sets the pace of AI advancement. Every insurer sets its own workforce readiness. The two futures where work turns out well are the two where readiness is widespread: people make the agentic leap in one, human and AI teams reshape value chains in the other. The two futures where it turns out badly are the two where readiness lags: entry pathways narrow in Stalled Progress, and reskilling systems simply cannot keep pace in the Age of Displacement.

The Forum's no-regret strategies for any scenario include aligning technology and talent strategies, designing multi-generational workflows, and investing in human-AI collaboration. Notice what those three share: each runs through the first rung. Workforce transformation is the readiness axis in motion, and readiness compounds the way judgement does: slowly, deliberately, from the bottom up.

Redesign the Rung: Where the Human-Agent Ratio Begins

The productivity story is real. Among entry-level workers, 68% report productivity gains from AI. Yet 45% also report spending more time working, and leaders describe freed capacity drifting into duplicative output and unreviewed machine drafts. One leader interviewed by the Forum called it what it is: "Without intentional job design, we're not creating value, we're just producing faster, lower-quality work, effectively creating 'work slop'."

The design gap is measurable. Only 16% of organizations report having fully redesigned roles, processes, and operating models to integrate AI effectively, while companies with the strongest AI-driven financial outcomes are twice as likely to redesign workflows rather than layer AI onto existing ones. Workforce transformation is not the compliance step after automation. It is the source of the return.

What does a redesigned first rung look like? The Forum's case studies offer insurance-ready patterns:

• Allianz pairs early career joiners with experienced colleagues in mentoring tandems: juniors bring agent fluency, veterans bring business context and judgement, and capability transfers in both directions.
• Dentsu Japan increased AI training time for new graduates more than tenfold and built a four-tier certification ladder, with 20,000 people already certified at the first tier. Progression became visible without waiting for tenure.
• Dropbox expanded its internship and graduate intake by 25% while automating aggressively, and reports that 87% of those early career hires earn promotion by year three, outperforming external hires.

In our work at Alchemy Crew, the design instrument is the Human-Agent Ratio: for every workflow, how much human attention it requires and where that attention creates the most value. A redesigned first rung sets that ratio deliberately, so juniors supervise, challenge, and exception-handle agent output from their first week. Human-in-command from day one, not after a decade of waiting.

From Theatre to Frontier: Index the Operating Model

Now connect workforce transformation to the operating model, because this is where insurers win or lose the decade.

The Evident AI Index for insurance (June 2026) found that 75% of disclosed AI outcomes across 30 major insurers are productivity gains and just 2% are revenue uplift, while only 8% of use cases qualify as end-to-end solutions that improve decision quality. Decision quality is the prize. And decision quality is a human-capability question before it is a model question: it depends on people who can judge, decide, and orchestrate.

This is why our Frontier Insurer Operating Model Index work benchmarks insurers across four tiers: Theatre, Experimenting, Scaling, Frontier. The instrument measures whether an operating model converts AI capability into business-model change, and workforce and human-agent design is one of its six pillars. It is also the pillar insurers most consistently underweight.

A carrier running AI theatre announces pilots and freezes graduate intake. A Frontier Insurer treats the first rung as infrastructure: intentionally designed, explicitly measured, and wired into how every one of the 10 emerging roles gets staffed 5 years from now. You cannot reach the Frontier tier with a hollowed pipeline, because every other pillar depends on the humans that pipeline produces.

Theatre automates the first rung and calls it efficiency. Frontier redesigns the first rung and calls it succession.

Five Moves for the Next Quarter

Workforce transformation earns its name only when it changes next quarter's decisions. Start with these five.

1. Put the first rung in the plan. Make entry-level intake an explicit line in strategic workforce planning, with targets that sit alongside AI adoption rather than behind it.
2. Redesign before you automate. For each entry-level role, answer three questions: where do humans create the most value, which tasks do we deliberately retain because they build capability, and does this role still develop what the organization will depend on?
3. Make progression visible. Replace tenure markers with capability markers. A third of entry-level workers plan to ask for promotion within a year; give them signals worth earning, Dentsu-style.
4. Pair the generations. Stand up mentoring tandems with agents in the middle: digital fluency flows up, judgement flows down, and people who understand both sides calibrate the ratio for each workflow.
5. Measure the conversion. Benchmark your operating model from Theatre to Frontier annually, and treat the workforce pillar score as a board metric, not an HR footnote.

The Cost of Standing Still

Do nothing for 90 days and little visibly changes. Do nothing for 3 years, and the arithmetic turns brutal: the graduates you did not hire in 2027 are the senior underwriters you will not have in 2035 and the chief underwriting officers you will not have in 2040. Agentic capability will be abundant, affordable, and everywhere. Seasoned judgement will not.

The agentic insurance enterprise will not be run by agents. It will be run by people who learned, earlier and faster than their predecessors, how to frame problems, judge evidence and command intelligent systems. Someone has to build the rung they climb first.

The real question is not whether your workforce transformation removes the first rung. It is whether you are the insurer deliberate enough to rebuild it.

Where does your first rung stand today? Ping me. 45 minutes. One conversation. A clear view of where your operating model and your talent pipeline sit on the road from Theatre to Frontier.

Frequently Asked Questions

Why are entry-level insurance jobs the most exposed to AI?

The World Economic Forum ranks financial and insurance activities first of 18 sectors for AI exposure because entry-level insurance work concentrates routine, information-processing tasks: data entry, document triage, standard assessments. Agentic systems execute these tasks most reliably today.

Will AI eliminate entry-level insurance jobs?

The evidence is mixed. Early career postings in high AI-exposure occupations have flatlined, yet the World Economic Forum still expects 170 million new jobs against 92 million displaced by 2030. Its Four Futures scenarios agree on the pivot: outcomes turn on workforce readiness, which insurers either build or abandon at the first rung.

What is the Human-Agent Ratio?

The Human-Agent Ratio is a per-workflow measure of how work divides between humans and AI agents: the agent-led share of a function's tasks, decisions, and volume, and where human attention creates the most value. It is the input metric for designing entry-level roles in an agentic enterprise.

How should insurers redesign entry-level roles for the agentic era?

Redesign around judgement rather than throughput: keep the tasks that build capability, put juniors in command of agent output from day one, pair them with experienced mentors, and make progression visible through capability markers instead of tenure. In any workforce transformation, role redesign precedes automation, never the reverse.

What is the Frontier Insurer Operating Model Index?

The Frontier Insurer Operating Model Index is an Alchemy Crew instrument that benchmarks whether an insurer's operating model converts AI capability into business-model change. It scores six pillars, including workforce and human-agent design, and bands insurers into four tiers: Theatre, Experimenting, Scaling and Frontier. 

Sources

• World Economic Forum and PwC, Artificial Intelligence and the Future of Entry-Level Work, June 2026
• World Economic Forum, Future of Jobs Report 2025, January 2025
• World Economic Forum, Four Futures for Jobs in the New Economy: AI and Talent in 2030, January 2026
• PwC, AI Jobs Barometer 2026 and Global Workforce Hopes and Fears Survey 2025
• PwC, 29th Annual Global CEO Survey, 2026
• Evident Insights, Evident AI Index: Insurance, Key Findings Report, June 2026