The Empty Desk Problem

AI Is Eliminating Entry-Level Busywork. Employers Must Reinvent Junior Training.

A commentary on AI, Gen Z, and the reinvention of professional training

There is a conversation happening in offices and firms across every professional sector right now. It goes something like this: "We use AI for the repetitive stuff now." The tone is usually upbeat. Efficient. Progressive. What that sentence quietly glosses over, however, is that the "repetitive stuff" has historically been the entire job description of every graduate who walked through the door in their first two years of professional life.

This article argues that AI's absorption of entry-level, process-driven work is not simply a workforce disruption story. It is also — if employers are brave enough to respond well — an overdue opportunity to reinvent how young professionals are actually trained. And it arrives at precisely the moment when Gen Z's disillusionment with traditional employment is already reshaping the labour market in ways that will be further accelerated by AI.

The Lie We Told Graduates

Here is what most planning, legal, financial, and consulting graduates are told, implicitly or explicitly, when they start work: put your head down, do the groundwork, learn the craft from the ground up, and the interesting stuff will come. What they actually receive is a keyboard, a library of old reports, and an instruction to use them as a template for the new one.

The copy-paste economy of junior work is not a secret. Reusing last year's environmental impact assessment structure for this year's site. Reformatting someone else's planning statement. Populating a spreadsheet that a partner will glance at for thirty seconds. The tasks are not without value in the abstract — they build familiarity with formats, processes, and industry language. But they provide almost none of the feedback, challenge, or meaning that transforms a graduate into a practitioner.

Many graduates leave those first years unable to answer a basic professional question: Am I any good at this? They have never been tested. They have only been processed.

"Gen Z workers consistently report that meaningful work, mentorship, and a sense of purpose matter more to them than any other workplace factor — yet the entry-level experience most firms offer provides almost none of these things."

What the Research Actually Says About Gen Z at Work

Gen Z's relationship with traditional employment was already under strain before AI entered the picture. The data is consistent and striking.

Research consistently shows that Gen Z prioritises meaningful work, mentorship, and clear career progression above almost everything else when evaluating an employer. Large corporations have recognised this: studies show Gen Z respondents prefer working at large organisations over small ones because they are better positioned to offer structured career paths, mentoring programmes, and competitive development opportunities.

— Accenture Strategy Report on Gen Z Rising; multiple Gen Z workplace preference surveys, 2022–2024

Yet the majority of entry-level roles in small and medium-sized professional firms — the ones that have traditionally absorbed new graduates — offer none of this infrastructure. They offer volume. And volume, it turns out, is exactly what AI does better.

Approximately 50% of Gen Z aspires to start their own business, and survey after survey shows increasing disillusionment with the nine-to-five as a model. This is not laziness — it is a rational response to an employment offer that fails to deliver what was promised.

— Multiple Gen Z entrepreneurship and career aspiration surveys, 2023–2024

The conditions were already there for a generational break with traditional employment pathways. AI is about to light a match.

What the Anthropic Economic Index Tells Us About Young Workers Right Now

A landmark study published in March 2026 by researchers Maxim Massenkoff and Peter McCrory at Anthropic introduces a new measure of AI displacement risk — "observed exposure" — that tracks not just what AI could theoretically do, but what it is actually doing in real professional settings.

The findings are sobering for anyone thinking carefully about workforce entry. The occupations with the highest current AI exposure — computer programmers (74.5% task coverage), customer service representatives (70.1%), data entry keyers (67.1%), and medical record specialists (66.7%) — are precisely the occupations that disproportionately absorb junior staff doing structured, repeatable work.

— Massenkoff & McCrory, "Labor Market Impacts of AI: A New Measure and Early Evidence," Anthropic, March 2026

More significantly, the study finds tentative but real evidence that hiring of young workers — specifically those aged 22 to 25 — has already begun to slow in AI-exposed occupations. In the post-ChatGPT era, the study records an approximately 14% drop in the job-finding rate for young workers entering highly exposed roles compared to 2022 baselines.

— Massenkoff & McCrory, 2026, Figure 7

This finding echoes parallel research by Brynjolfsson, Chandar, and Chen (2025), which reports a 6–16% fall in employment in AI-exposed occupations among workers aged 22 to 25, attributing the decrease primarily to a slowdown in hiring rather than job losses among existing workers.

— Brynjolfsson et al., "Canaries in the Coal Mine? Six Facts About the Recent Employment Effects of Artificial Intelligence," Digital Economy, 2025

The canary in the coal mine, to borrow Brynjolfsson et al.'s evocative framing, is singing. The question is whether employers are listening — and if so, what they plan to do about it.

"The occupations most exposed to AI automation are, almost without exception, the ones that have historically served as the training ground for professional life. That is not a coincidence. It is a structural problem."

The Gap Between Capability and Deployment — and Why It Matters

One of the most important findings in the Anthropic report is the gap between what AI can theoretically do and what it is currently doing in practice. Despite LLMs being theoretically capable of covering 90% of tasks in office and administrative occupations, actual observed coverage sits far below that ceiling — and the gap is narrowing.

— Massenkoff & McCrory, 2026, Figure 2

This means the displacement of entry-level work is not a future scenario. It is an ongoing, accelerating process. The tasks that graduates are currently being asked to do — reformatting documents, populating templates, producing first drafts of standardised reports — are being automated at pace. And yet the training models built around those tasks have not changed.

Firms that continue to onboard graduates as if AI does not exist are not preserving a valuable apprenticeship model. They are handing graduates a disappearing job description and calling it development.

The Reinvention That Is Now Required

If the entry-level busywork is going — or has already largely gone — then the organisations that survive this transition well are the ones that ask a harder question: what does a junior professional actually need to learn, and how do we teach it without the scaffolding of repetitive tasks?

This is genuinely difficult. The volume model of junior training, whatever its flaws, did produce some real benefits: familiarity with document structures, exposure to a wide variety of project types, a slow-burn absorption of how a firm's work is organised. Replacing it requires intentionality. What might that look like?

Structured mentorship paired with AI review. Juniors work alongside AI tools to produce outputs, but are explicitly trained to critique, interrogate, and improve AI-generated work rather than simply submit it. This builds judgment rather than just output.

Early exposure to client interaction and problem framing. The tasks AI cannot easily replicate — understanding a client's actual concern, navigating ambiguity, building trust — can be introduced earlier when repetitive work is no longer consuming all of a junior's time.

Real feedback loops from day one. One of the most damaging features of traditional junior work is the absence of meaningful feedback. Did I do this well? What did the partner actually think? AI-assisted workflows can free up senior staff time for the kind of deliberate coaching that builds competent professionals faster.

Portfolios of judgment, not outputs. Rather than measuring a junior's first year by the number of reports filed, firms could measure the quality of decisions made, alternatives considered, and problems identified — skills that are both more valuable and more transferable than template-filling.

The Brave Employers: Taking On Entry-Level Staff in the Age of AI

There will be an easy path and a hard one for employers over the next decade.

The easy path is to stop hiring juniors altogether. If AI can do the work, why pay a graduate to do it more slowly and with more mistakes? This logic is already being acted upon, as the hiring data suggests. It is short-sighted, but understandable.

The hard path is to hire juniors anyway, and to do the genuinely difficult work of reinventing how they are developed. This requires investment in mentorship infrastructure, willingness to redesign workflows, and a long-term view of what a firm needs to look like in ten years.

The employers who take the hard path will not just be doing something ethically important — though they will be. They will be building something that the firms who took the easy path will eventually need to buy back at a premium: a pipeline of professionals with genuine judgment, not just operational competence.

"The firms that take on entry-level staff in the age of AI are not being naive. They are making a long bet on human development at the precise moment everyone else has decided it is too complicated."

This matters especially for Gen Z. A generation that already approaches traditional employment with scepticism, that already leans toward entrepreneurship and autonomy, is not going to be retained by organisations that offer them a front-row seat to watching AI do their job. They will stay — and thrive — in organisations that treat their early years as a genuine investment in professional formation.

Research is already clear that small and medium-sized businesses are poorly positioned to meet this challenge. The infrastructure required — structured mentorship programmes, deliberate onboarding frameworks, meaningful feedback systems — is resource-intensive. This creates a structural advantage for larger organisations willing to invest, and a real risk that the gap between firms that develop talent well and those that do not will widen significantly.

— Multiple SMB workforce research reports, 2023–2024

A Prediction

The Gen Z disillusionment with entry-level work that is already visible in the data will be significantly exacerbated by AI's absorption of the tasks that have defined those roles. The combination of a generation that already found junior work unrewarding and a technology that is now replacing the most automatable parts of it will accelerate departures from traditional employment pathways.

Firms that respond to this by quietly reducing graduate intake will not feel the consequences immediately. They will feel them in five to eight years, when the professionals who would have been their senior associates and directors either do not exist, have gone elsewhere, or have built their own practices.

The Anthropic Economic Index will continue to update. The gap between AI's theoretical capability and its observed workplace coverage will keep narrowing. The occupations being eroded are not mysterious — they are listed in plain sight, and they are the occupations that train the next generation of every profession that depends on knowledge work.

The question for every firm reading this is simple: are you going to reinvent how you develop people, or are you going to wait and see? Because the graduates you are not hiring right now are the senior professionals you will be trying to recruit in a decade.

Key Sources

• Massenkoff, M. & McCrory, P. (2026). Labor Market Impacts of AI: A New Measure and Early Evidence. Anthropic.
• Brynjolfsson, E., Chandar, B. & Chen, R. (2025). Canaries in the Coal Mine? Six Facts About the Recent Employment Effects of Artificial Intelligence. Digital Economy.
• Accenture Strategy. Gen Z Rising: The Next Generation of Talent. Accenture.
• Multiple Gen Z workplace preference and entrepreneurship surveys (2022–2024).
• SMB workforce research reports on Gen Z hiring infrastructure gaps (2023–2024).