AI Will Not Replace Your Operating Model. It Will Expose It.

AI discussions often start with jobs.

Which roles will disappear?

Which skills will become obsolete?

How many people will we need?

These are reasonable questions.

But they start too late.

The first question should be:

How will the work change?

Because when AI changes how work gets done, the operating model changes with it.

Roles change.

Decision rights change.

Teams change.

Skills change.

The boundaries between functions change.

AI does not arrive in an organisation with a blank operating model.

It enters the one already there.

And exposes its weaknesses.

Start with the work

Consider a process where people spend significant time gathering information, producing reports or responding to routine enquiries.

AI might reduce the effort involved.

But the value does not come from removing the task alone.

The organisation needs to decide what happens next.

Does the person take on more complex work?

Does decision-making move closer to the customer?

Does the role require stronger judgement?

Does another team now own part of the process?

Does the process itself need redesign?

Automation changes the economics of work.

AI changes the shape of work.

The operating model needs to respond to both.

AI exposes organisational boundaries

Many enterprise problems sit between functions.

A customer issue might involve sales, operations, technology, finance and service.

The process crosses organisational boundaries even though the organisation chart does not.

AI makes this more visible.

If the problem is cross-functional, the people solving it need to work across functions too.

Enterprise AI therefore requires more than technical capability.

It requires multidisciplinary collaboration.

The traditional model of deep expertise in one discipline, supported by broader capability, starts to evolve.

This evolving capability is sometimes described as the M-shaped employee: expertise across multiple connected areas, supported by integrated knowledge across disciplines.

The important point is not the letter.

The important point is the shift in capability.

The future workforce needs more than AI skills

A common response to AI is to launch training.

Teach people how to use the tools.

Teach prompt techniques.

Teach new platforms.

Useful, but insufficient.

AI tools will continue to change.

People need the ability to adapt alongside them.

Curiosity.

Judgement.

Interdisciplinary thinking.

Continuous learning.

The ability to question assumptions.

The confidence to work through ambiguity.

The ability to understand where human judgement matters most.

Continuous learning needs to become part of work itself rather than remain an occasional training intervention.

This changes the role of leadership too.

Leaders need enough confidence to make decisions without having every answer upfront.

They need to encourage experimentation.

They need to create space for people to challenge assumptions.

And they need to recognise when previous knowledge is no longer sufficient.

Human capability becomes more important

There is a temptation to think about AI in terms of what humans no longer need to do.

A better question is:

What should humans do more of?

AI is well suited to labour-intensive analysis, information processing and repetitive activity.

People bring judgement.

Creativity.

Relationship management.

Empathy.

Interpersonal problem-solving.

Context.

The ability to deal with situations where the answer is not obvious.

Examples include AI supporting creative testing, scenario-based forecasting and customer service while people retain responsibility for creativity, judgement and complex problem-solving.

The opportunity is not human versus AI.

The opportunity is to redesign the relationship between them.

The operating model has to move

This is where many AI programmes become disconnected from transformation.

The technology team deploys the capability.

HR updates the skills framework.

Learning launches training.

The business identifies use cases.

Risk develops controls.

Each function does its part.

But the work itself does not change coherently.

The result is another layer added to the existing operating model.

AI adoption without operating model change creates friction.

People use new tools inside old processes.

New capabilities sit inside old decision structures.

New roles inherit old responsibilities.

The organisation gets AI activity without enough business value.

Ask different executive questions

Instead of asking:

How many employees will AI replace?

Ask:

Which work should AI perform?

Then ask:

Which work should humans perform?

Where does judgement need to stay with people?

Which roles need broader capability?

Which decisions should move closer to the work?

Which organisational boundaries are now getting in the way?

What new skills need to become part of everyday work?

Which existing roles need to be redesigned rather than removed?

These questions move the conversation from workforce reduction to work redesign.

That is a much more consequential conversation.

AI is an operating model intervention

The real AI transformation is not the introduction of another technology.

It is the redesign of how work gets done.

That includes:

Work

What tasks change?

Roles

What do people own after AI takes on more activity?

Teams

Where does multidisciplinary collaboration become necessary?

Decision rights

Which decisions stay with people?

Which decisions move closer to AI?

Capabilities

What combination of technical, business and human skills is required?

Learning

How does the organisation keep people current as AI evolves?

Leadership

How do leaders create the conditions for experimentation, judgement and adaptation?

These are operating model decisions.

They belong in the transformation conversation.

The executive question is not about headcount

AI will change workforce requirements.

Some roles will shrink.

Some will expand.

New roles will emerge.

Existing roles will be redesigned.

But headcount is an output of those changes.

It should not be the starting point.

The starting point is the business outcome.

Then the work required to deliver it.

Then the operating model required to perform that work.

Then the capabilities people need.

Then the technology that enables the model.

This sequence matters.

Otherwise the organisation risks using AI to optimise yesterday's operating model.

The real test

The question for leadership is not:

How many people will AI replace?

It is:

What should our organisation look like when AI becomes part of how work gets done?

That question forces a broader conversation.

About work.

About accountability.

About skills.

About leadership.

About organisational boundaries.

About decision-making.

About how humans and AI work together.

AI will not replace your operating model.

It will expose whether your operating model was designed for the work you need to do next.