WHO WILL BECOME SENIOR?

The work executives want to eliminate may be the work that teaches people how the business works

Imagine a professional-services firm reviewing its AI opportunities.

Junior analysts spend thousands of hours researching industries, reviewing documents, preparing financial analyses, drafting reports and building presentations.

AI can perform substantial portions of this work.

The business case looks compelling.

Reduce junior hiring.

Increase output per professional.

Improve margins.

Then ask a different question:

Where did today’s partners learn to become partners?

Often, they learned by doing exactly the work the firm is preparing to automate.

That should change the conversation.

Junior work has two outputs

Organizations usually measure junior work by its immediate output.

The analyst produces an analysis.

The accountant prepares a working paper.

The engineer completes a calculation.

The junior lawyer researches a case.

But the work produces something else:

a more experienced employee.

That second output rarely appears in the business case.

Suppose AI performs a junior employee’s task for one-tenth the cost.

The immediate productivity calculation is obvious.

But if performing that task contributed to developing pattern recognition, commercial judgment or professional intuition, the long-term calculation becomes considerably more complicated.

The hidden apprenticeship system

This issue extends far beyond professional services.

Junior employees learn how customers behave.

Supervisors learn how operations fail.

Analysts learn what good data looks like.

Managers learn how people respond under pressure.

Salespeople learn which objections actually matter.

Much of this knowledge is tacit.

It is accumulated through exposure rather than formal training.

AI can transfer information remarkably well.

Experience is harder.

The CEO question

Before automating a significant category of junior work, ask:

What capability was this work unintentionally teaching?

Then ask:

How will we develop that capability after the work disappears?

If management cannot answer the second question, the automation business case is incomplete.

The Arcus perspective

Workforce planning in the AI era cannot simply forecast how many employees are required.

It must identify how organizational capability is created over time.

An Arcus Future Workforce Architecture Review can map critical roles, career pathways, learning experiences and AI exposure to determine where today’s productivity decisions may create tomorrow’s capability gaps.

The objective is not to preserve unnecessary work.

It is to preserve the learning that the work used to provide.

Consider a hypothetical consulting firm.

It traditionally hires 100 graduates every year.

Ten years later, perhaps 15 have developed into senior professionals.

AI allows the firm to reduce graduate recruitment from 100 to 30.

The immediate economics improve.

But ten years later, where do the 15 senior professionals come from?

This is the arithmetic behind one of AI’s least discussed organizational risks.

The pyramid can shrink from the bottom

Many organizations assume AI will flatten their hierarchy.

That may happen.

But flattening the pyramid from the bottom creates a delayed effect.

You do not see the leadership shortage immediately.

You see it years later.

This is why conventional workforce planning may miss the problem.

The pipeline equation

A simple workforce model should connect:

Entry population × development rate × retention rate × time = future senior capability.

If any component changes materially, future supply changes.

AI is likely to alter all four.

Organizations may recruit fewer junior employees.

Employees may develop differently.

Career expectations may change.

High performers may leave if progression becomes unclear.

The risk is concentrated

The problem could be particularly acute in organizations where expertise takes years to develop:

• engineering;
• banking;
• law;
• consulting;
• accounting;
• medicine;
• insurance;
• manufacturing;
• infrastructure;
• government;
• technical industries.

A senior engineer cannot simply be manufactured through an online course.

Neither can an experienced commercial banker or plant manager.

What CEOs should request

Ask HR to show the executive team not simply next year’s workforce plan but a five- and ten-year capability pipeline for the organization’s most critical roles.

Where do those people currently come from?

How long does development take?

Which developmental experiences are being automated?

What happens if entry-level hiring falls by 30, 50 or 70 percent?

The answers may change today’s hiring decisions.

There is a compelling argument for reducing entry-level hiring.

AI can increasingly perform the work.

But there is an equally compelling argument for keeping junior employees.

They are inexpensive relative to senior talent.

They bring new skills.

They become future leaders.

So how should management reconcile the two?

The answer is not preserving obsolete jobs.

It is redesigning entry-level work.

The traditional bargain is breaking

For decades, many junior professional roles operated under an implicit bargain:

You will perform relatively routine work today because doing it teaches you how to perform valuable work tomorrow.

AI changes the first half.

Organizations therefore need to redesign the second.

The junior employee after AI

Instead of spending 20 hours producing an analysis, a junior employee may spend two hours using AI to produce it.

What happens during the remaining 18?

That is the strategic opportunity.

The employee could:

• test assumptions;
• interview customers;
• observe senior decision-making;
• participate in client discussions;
• evaluate AI outputs;
• explore alternative scenarios;
• work across functions;
• conduct experiments.

The role can become more developmental, not less.

But that requires intentional design

Organizations should not assume this happens automatically.

Without intervention, the saved time will be absorbed by additional work.

The organization may produce twice as many analyses while developing half as much judgment.

The CHRO opportunity

This may be one of the rare moments when technology allows organizations to improve both productivity and human development.

But only if job architecture changes with the technology.

The objective should not be to preserve the entry-level job of 2020.

It should be to design the entry-level job of 2030.

How do you learn to do work that AI already does for you?

Imagine a young financial analyst.

AI constructs the model.

It identifies anomalies.

It generates scenarios.

It drafts the investment memorandum.

The analyst reviews the result.

That sounds highly productive.

But how does the analyst know whether the model is wrong?

That question exposes the experience paradox.

Expertise requires pattern recognition

Experienced professionals often recognize problems before they can fully articulate why.

A number looks wrong.

A customer explanation feels incomplete.

A contract clause creates concern.

A production sound indicates trouble.

This is not magic.

It is accumulated exposure.

Thousands of previous situations create mental patterns.

AI can remove exposure

If AI increasingly handles the first pass through the material, employees encounter fewer raw problems.

They see polished answers rather than messy inputs.

That can accelerate output while slowing learning.

We have seen analogous problems elsewhere.

Automation in aviation dramatically improves safety, but pilots still require opportunities to develop and maintain skills for situations where automation fails.

Organizations may face a similar challenge with knowledge work.

The new learning architecture

Future employee development may need deliberate mechanisms for creating exposure:

• simulations;
• case reviews;
• rotations;
• shadowing;
• supervised decision-making;
• scenario exercises;
• post-mortems;
• AI-output challenges;
• apprenticeship with senior employees.

Experience can no longer be assumed to emerge naturally from workload.

It may need to be engineered.

A powerful management question

Ask senior employees:

What experiences made you good at your job?

Then identify which of those experiences junior employees are likely to lose because of AI.

That gap should become part of the workforce strategy.

Middle management has become an attractive target for organizational simplification.

The logic appears compelling.

AI can prepare reports.

Dashboards provide information directly to executives.

Agents can coordinate tasks.

Communication can bypass management layers.

Therefore, remove managers.

Sometimes that will be correct.

But there is a hidden problem.

Managers perform invisible work

A good middle manager does far more than transmit information.

They notice when a strong employee is struggling.

They explain why senior management made an unpopular decision.

They resolve disagreements before they become organizational conflicts.

They teach new employees how the company really works.

They recognize talent.

They interpret culture.

They know which customer is actually unhappy despite the dashboard showing green.

These activities rarely appear in job descriptions.

That makes them easy to eliminate accidentally.

Remove roles, not functions

Before eliminating a management layer, identify every function it performs.

Then divide those functions into:

Administrative

AI may perform these exceptionally well.

Coordination

AI may perform parts of these.

Decision

AI may augment these.

Human

Coaching, trust, conflict, motivation and development may become more important.

The paradox

AI may allow organizations to employ fewer managers while requiring those remaining managers to become considerably better leaders.

That has implications for selection, compensation and training.

The manager who derived authority from controlling information may become less valuable.

The manager who creates capability in other people may become more valuable.

The CEO question

Do not ask only:

“Can this management layer be removed?”

Ask:

“What organizational capability disappears with it?”

Then decide deliberately whether AI, another role or a redesigned manager should provide that capability.

It sounds contradictory.

AI is supposed to reduce labour shortages.

Yet one possible consequence is a shortage of experienced leaders.

The reason is time.

Leadership has a long production cycle

A strong executive may represent 20 years of accumulated experience.

During those years the person has:

• managed difficult employees;
• lost customers;
• recovered from mistakes;
• handled crises;
• allocated budgets;
• negotiated internally;
• made unpopular decisions;
• watched strategies fail;
• learned an industry.

AI can provide information about these experiences.

It cannot give someone twenty years of having lived them.

The delayed shortage

If organizations reduce junior recruitment today, narrow management roles tomorrow and automate developmental work throughout the hierarchy, the consequences may remain invisible for years.

Then organizations begin competing for a shrinking population of experienced people.

The irony would be striking.

AI lowers the cost of analytical intelligence while increasing the scarcity premium attached to human judgment.

This changes succession planning

Traditional succession planning asks:

Who could replace this executive?

Future succession planning should also ask:

What developmental experiences must exist across the organization so that somebody is capable of replacing this executive five years from now?

That is a much harder question.

Build capability inventories

Organizations track financial assets meticulously.

They should begin treating critical human experience similarly.

Where does the organization possess scarce judgment?

How old is that population?

How quickly can it be replaced?

Which individuals hold institutional knowledge?

Where is there no credible successor?

Which developmental pathways are weakening?

Those questions belong on the executive agenda.

One of the most significant changes AI creates is deceptively simple.

Junior employees may increasingly receive completed work rather than blank pages.

The job shifts from:

produce

to:

prompt, evaluate, challenge, improve and decide.

That sounds like progress.

But reviewing good work requires knowledge.

The review paradox

A senior lawyer can recognize a weak AI-generated argument because the lawyer understands the law.

A senior engineer can detect an unrealistic recommendation because the engineer understands the system.

A first-year employee may not.

This creates an unusual problem:

AI can allow inexperienced employees to produce work that looks more sophisticated than their ability to evaluate it.

Fluency can masquerade as competence.

Organizations need verification capability

Training therefore needs to move beyond:

How do you use AI?

toward:

How do you know when AI is wrong?

That requires:

• domain knowledge;
• source verification;
• reasoning;
• statistical literacy;
• professional standards;
• skepticism;
• judgment.

Ironically, some traditional foundational knowledge may become more important rather than less.

A new competency model

Organizations should begin evaluating employees across three dimensions:

Production capability

Can the employee create high-quality work?

AI orchestration capability

Can the employee use AI effectively?

Judgment capability

Can the employee recognize when the answer should not be trusted?

The third may ultimately become the most valuable.

A twenty-three-year-old employee may be substantially better at using AI than a fifty-five-year-old executive.

That does not necessarily mean the junior employee is better equipped to make the decision.

Conversely, decades of experience do not guarantee that a senior executive understands how AI changes the economics of the business.

Organizations therefore face a two-way capability gap.

Younger employees may possess technological fluency

They experiment quickly.

They adopt new tools.

They discover workflows.

They challenge historical processes.

Experienced employees possess contextual capital

They understand:

• customers;
• organizational politics;
• regulatory boundaries;
• historical failures;
• commercial trade-offs;
• reputational consequences.

The organization needs both.

Reverse mentoring is not enough

The opportunity is deeper than teaching executives how to prompt.

Organizations can deliberately pair:

AI-native capability + experienced judgment.

Imagine cross-generational teams where younger employees redesign workflows while senior employees test assumptions and provide context.

Both groups learn.

The workforce advantage

Companies frequently treat generational differences as an HR problem.

AI could turn them into a strategic asset.

The organizations that combine technological fluency with accumulated judgment may outperform those that simply replace one generation’s capabilities with another’s.

Traditional careers were comparatively linear.

Analyst.

Manager.

Director.

Vice-president.

Executive.

Each promotion brought larger teams, broader responsibility and greater compensation.

AI may weaken this model.

If organizations become flatter and individual contributors become dramatically more leveraged, managing more people may no longer be the primary route to greater value.

Expertise can scale differently

An exceptional specialist supported by AI may create more enterprise value than a manager supervising twenty employees.

Organizations therefore need career paths that reward:

• deep expertise;
• problem-solving;
• customer impact;
• intellectual property;
• innovation;
• AI orchestration;
• enterprise contribution.

Without them, talented employees may continue seeking management positions simply because that is where status and compensation reside.

The career network

Future careers may involve movement between:

specialist → project leader → business role → expert → team leader → enterprise role.

Progress becomes accumulation of capability rather than movement up a single hierarchy.

Why this matters now

If organizations flatten structures without redesigning careers, employees will perceive fewer opportunities.

That creates a retention problem precisely when scarce capability becomes more valuable.

The answer is not preserving unnecessary hierarchy.

It is separating career progression from management hierarchy.

The CHRO agenda

Compensation, titles, development and succession systems all need to adapt.

The organization that redesigns work but leaves the career architecture untouched has completed only half the transformation.

The central workforce question created by AI is not:

How many jobs will disappear?

Nobody can answer that reliably.

A more useful question is:

What human capabilities will this organization need ten years from now, and how are we going to produce them?

That is answerable.

And it should become part of strategy.

Start with capabilities, not headcount

Take the organization’s 20 most strategically important capabilities.

For each determine:

1. What human judgment does it require?
2. How long does that judgment take to develop?
3. Which roles currently produce it?
4. Which developmental experiences create it?
5. Which of those experiences will AI automate?
6. Where could AI accelerate development?
7. What happens if junior hiring declines?
8. Where is external recruitment realistic?
9. Where is experience genuinely scarce?
10. Who owns the pipeline?

The result may look very different from a conventional workforce plan.

The board should see this

Boards routinely examine:

• capital requirements;
• technology investments;
• succession;
• strategic risks.

The long-term production of organizational capability deserves similar attention.

A company can purchase technology.

It can acquire equipment.

It can raise capital.

Some forms of experience cannot be purchased quickly at any price.

A new workforce architecture

Arcus believes the workforce implications of AI should be examined across four interconnected layers:

WORK

What work should humans and AI perform?

ROLES

How should jobs change?

CAREERS

How will employees acquire increasingly valuable capabilities?

PIPELINE

How will the organization produce the leaders and experts it will require five and ten years from now?

This moves workforce planning from an HR forecasting exercise into enterprise strategy.

The Arcus Future Workforce Architecture Review

For organizations confronting significant AI-driven workforce change, Arcus can work with the CEO, CHRO and leadership team to examine:

• AI exposure by role;
• critical capability dependencies;
• job and role redesign;
• entry-level workforce strategy;
• management structure;
• spans and layers;
• career architecture;
• learning and apprenticeship;
• leadership pipelines;
• succession vulnerabilities;
• five- and ten-year workforce scenarios.

The objective is not to predict exactly how many employees AI will replace.

It is to avoid a much more dangerous outcome:

discovering several years from now that today’s efficiency decisions eliminated the experiences that created tomorrow’s leaders.