The People You Remove Today May Be the People You Need Tomorrow

There was a time when the accepted response to economic disruption was to protect capability. Organisations were encouraged to use slower periods to train employees, improve processes and prepare for recovery. The argument was commercially straightforward: skills, customer knowledge and institutional memory had taken years to build and could not be reconstructed immediately when demand returned.

That principle has not disappeared. It has, however, been challenged by a more aggressive proposition. If artificial intelligence can absorb administrative work, answer routine questions, analyse data and generate content, an organisation may conclude that it no longer requires the same number of people.

The proposition is attractive because payroll is visible, immediate and measurable. The commercial value of judgement, informal coordination, customer trust and institutional knowledge is considerably harder to express in a board paper.

These decisions do not necessarily arise from indifference. Executives may be responding to investor expectations, margin pressure, technological change and competitive threats. A board that believes competitors will achieve the same output with fewer people may consider inaction a greater risk than restructuring.

The analytical weakness appears when the organisation compares the cost of a person with the cost of a system, but fails to compare the full capability of each.

Customer impact

Many customers are comfortable using automated systems for simple, low-risk tasks. A capable tool available immediately may be preferable to waiting for a person. The problem begins when the system cannot recognise the limits of its own competence.

Customers increasingly encounter automated channels that repeatedly answer the wrong question, prevent escalation or require the customer to repeat information. The organisation may record a reduction in average handling cost while the customer experiences a transfer of work from the company to themselves.

Employee impact

The effect extends beyond those who leave. Remaining employees observe how colleagues are treated and draw conclusions about the relationship between productivity and security.

An organisation cannot credibly ask employees to experiment with AI while implying that every successful efficiency gain may justify another reduction. Under those conditions, employees may withhold knowledge, avoid visible experimentation or use technology defensively rather than creatively.

Capability impact

Removing a role does not guarantee that the underlying capability disappears from the organisation's needs. Tasks often move to managers, customers, contractors or surviving teams. The apparent cost saving may therefore become hidden labour elsewhere.

There is also a succession risk. Junior roles often contain repetitive work, but they also provide the training ground through which employees learn judgement. If the organisation automates the entry route without redesigning development, it may preserve today's experts while eliminating the pathway that creates tomorrow's experts.

Financial implications

Redundancy can produce a visible reduction in operating cost. The longer-term financial calculation must also consider recruitment, onboarding, retraining, customer churn, complaint handling, lost productivity, slower decision-making and the cost of rebuilding relationships.

A capability removed in one financial year may return as a recruitment requirement in the next. Rebuilding it can be more expensive because former employees may have moved on, salary expectations may have changed and trust in the organisation may have weakened.

The debate is often presented as a contest between people and technology. That framing is too simplistic.

Organisations that reject capable technology will become less competitive. Organisations that remove people without understanding the capability contained within their roles may become more efficient on paper and less effective in practice.

The central leadership task is therefore to design the division of responsibility. AI should provide speed, analysis, consistency and scale. People should retain ownership where trust, complexity, judgement and consequence are highest.

The people closest to the work should be involved in the redesign. Employees often understand process failure, customer exceptions and data limitations better than the executives purchasing the system. Their involvement improves the quality of implementation and reduces the risk that important work is simply made invisible.

The commercial objective should not be to preserve every role exactly as it exists. It should be to preserve and develop the capabilities the organisation will require as roles change.