AI Is an Amplifier. What Is It Amplifying?

When I first started thinking seriously about AI, I saw it as a catalyst.

A catalyst accelerates a reaction.

But it does not create the underlying material.

Over time, another word became increasingly important to me:

Amplifier.

Because AI does more than accelerate.

It increases the reach, speed and intensity of what is already there.

Your knowledge.

Your curiosity.

Your creativity.

Your judgment.

But also your assumptions.

Your impatience.

Your confusion.

And that raises a more interesting question than:

What can AI do?

What is AI amplifying?

The source still matters

In Modern Alchemy, I described a simple principle:

Use AI to amplify, not replace.

The experience, intuition, questions, doubts and convictions have to come from somewhere.

The human remains the source.

My shorthand for working with AI became:

Clean. Structure. Amplify.

Give AI something meaningful and it can create extraordinary leverage around it.

Give it confusion and it can produce extremely sophisticated confusion.

That principle matters even more as AI becomes more capable.

Because a more capable AI does not automatically create a more capable leader.

It gives the leader more leverage.

And leverage works in both directions.

Amplification changes the scale of the consequences

Consider a leader who has reached a conclusion too quickly.

Before generative AI, turning that conclusion into an organisational initiative required friction.

Research had to be gathered.

Arguments constructed.

Documents written.

Presentations prepared.

Questions answered.

That friction wasn't always efficient.

But it created time.

Today, the same leader can turn an initial assumption into a compelling strategic memo, supporting analysis, an executive presentation and an implementation plan in a fraction of the time.

The original assumption has not necessarily become better.

Its ability to travel has.

That is the difference between the mirror and the amplifier.

The question is no longer only whether AI can help us recognise a weak assumption.

The question becomes:

What happens when a weak assumption can be industrialised and propagated across an organisation at unprecedented speed?

The amplification is not symmetrical

This is where I think the conversation about AI and productivity becomes too narrow.

A leader's operating system can be described as a loop:

Learn → Think → Decide → Act → Observe → Adjust

AI can contribute across that entire loop.

But it does not accelerate every part equally.

It can dramatically accelerate access to information and therefore parts of Learn.

It can expand Think, provided we don't outsource thinking to it.

It can support Decide, without taking responsibility for judgment.

And it can massively accelerate the production that follows a decision:

the memo,

the presentation,

the communication,

the plan,

the analysis,

the code.

But producing the artefacts of action is not the same as changing reality.

A presentation can be generated in minutes.

An organisation still needs time to understand it.

A strategic narrative can be rewritten before lunch.

People still need time to believe it.

A new operating model can look beautiful on a slide.

Changing behaviour takes longer.

AI accelerates representation of change far more easily than change itself.

That asymmetry matters.

AI can make the leader faster than the organisation

Imagine a CEO exploring a new strategic direction before breakfast.

By 9am, AI has helped challenge and develop the idea.

By 10am, there is an executive memo.

By 11am, there is a compelling deck.

By noon, a new priority can be communicated.

From the CEO's perspective, this can feel like extraordinary strategic agility.

But the organisation is operating on a different clock.

People are still implementing yesterday's priorities.

Teams are still interpreting the previous decision.

Customers are still responding to changes made months earlier.

Reality has not accelerated simply because the leader has.

Then, two days later, another compelling idea emerges.

The cycle starts again.

Each individual decision may even be reasonable.

Yet collectively they create turbulence.

AI can increase the speed of the leader faster than the organisation's ability to absorb the consequences.

At that point, velocity stops being an advantage.

It becomes a leadership problem.

The ability to accelerate creates a new responsibility: knowing when not to

Much of the current conversation around AI asks how leaders can move faster.

That is understandable.

Speed creates competitive advantage.

But once speed becomes abundant, restraint becomes more valuable.

The question changes from:

How can I accelerate this?

to:

Should this be accelerated?

Does the organisation need another answer?

Another initiative?

Another communication?

Another strategic adjustment?

Or does it need enough time for the previous decision to meet reality?

The ability to generate change faster increases the responsibility to regulate its frequency.

The scarce leadership capability may no longer be acceleration. It may be calibration.

Observe may become more important, not less

This brings me to the part of the operating system I find increasingly important:

Observe.

AI can analyse surveys.

Summarise meetings.

Detect patterns.

Compare performance.

Model scenarios.

All useful.

But observation is not simply data collection.

Sometimes it means noticing that nobody challenged the idea in the room.

Sensing that an organisation is tired.

Discovering that a strategy everyone understood intellectually has not translated into behaviour.

Listening to what customers actually say.

Walking through an operation.

Giving a decision enough time to encounter reality.

Operational reality has its own clock.

And the faster AI accelerates Learn → Think → Decide → Produce, the more deliberate leaders may need to become about Observe → Adjust.

Otherwise the loop changes.

It stops being a learning system.

It becomes an acceleration system.

And those are not the same thing.

The amplifier needs calibration

This is where my thinking about AI has evolved.

I still see AI as a catalyst for change.

I still believe it can act as a powerful mirror.

I deliberately introduce friction when I want it to challenge rather than reinforce my thinking.

But amplification introduces another dimension.

Scale.

A thought becomes an analysis.

An analysis becomes a decision.

A decision becomes communication.

Communication becomes organisational movement.

AI can compress the time between those stages dramatically.

The challenge for leaders is therefore not simply to become faster with AI.

It is to ensure that their feedback loop becomes stronger as their acceleration increases.

That is a very different definition of AI maturity.

So what is AI amplifying?

Your curiosity?

Your judgment?

Your ability to learn?

Your ability to make decisions?

Or simply your ability to produce?

And beyond yourself:

What is that amplification doing to the system around you?

Is it creating movement?

Or turbulence?

Is the organisation learning faster?

Or simply receiving more instructions?

Are you observing reality?

Or already generating the next answer?

AI gives leaders extraordinary leverage.

The interesting question is no longer whether we can accelerate.

We can.

The question is whether our judgment, our feedback loops and our organisations can absorb the speed we are creating.

AI is an amplifier.

Leadership determines the signal — and the gain.