
The Tide: Why AI Is Infrastructure, Not a Tool

Marc Lewis, Head Coach at the School of Communication Arts 2.0 in London, wrote this talk over the winter holidays in 25/26 and shared it to the dutch advertising and publishing community during 'the Agentic Leadership Journey: The Reunion' on Tuesday January 13th . Between courses, guests were invited to discuss three questions:
- What are we at risk of losing by offloading too much to AI?
- If AI frees creative capacity, what are we deliberately directing that capacity towards?
- If we rebuilt our agency around today’s reality, what would we not design back in?
Good evening.
Thank you for having me.
A particular thank you to Brian Hirman Mirjam Assink and the team at VIA Nederland for making this possible.
I want to start with a confession.
I am a tech optimist.
But I am a tech optimist who has been burned.
Which is a polite way of saying I’ve been wrong, in public.
Expensively.
But in the late 90s, I was part of the first wave of so-called digital prophets.
I built a start-up.
I lived on adrenaline.
And I sold it to Sir Martin Sorrell and Lord Bell for twenty million pounds…
…just as the clock struck midnight on the dot-com bubble.
I’ve seen what happens when the New becomes Normal.
And I’ve seen what happens to the people who mistake a temporary bubble
for a permanent shift.
I’ve spent decades since then nurturing creative talent.
Helping people win Lions.
Helping people win Pencils.
And more importantly
helping them lead.
In all that time, I have never seen a moment as misunderstood as this one.
When you walk into your agencies tomorrow, listen to how people talk about AI.
They talk about tools.
Midjourney.
Sora.
ChatGPT.
They argue about “prompt engineering” as if it were a career path.
They obsess over efficiency.
How fast can we churn out a storyboard?
How cheap can we make a social post?
This is the wrong conversation.
Focusing on AI as a tool
is like standing on a beach, staring at one impressive wave
and missing the fact that the tide has turned.
Tonight isn’t about the waves.
It’s about the tide.
Because the most important technologies in history
don’t arrive as gadgets.
They arrive as infrastructure.
And infrastructure doesn’t just make things faster.
It rearranges the world
while we aren’t looking.
Since we’re in one of the world’s great port cities,
let’s talk about the most boring revolution in history.
The shipping container.
Before 1956, global trade was chaotic, human, and deeply expert.
If you walked onto the docks in Amsterdam,
it was a theatre of skill.
Men knew how to stow grain next to cloth so neither spoiled.
They knew how to balance a ship by feel.
It was physical.
It was dangerous.
And it was fiercely local.
Then came Malcolm McLean.
He wasn’t a sailor.
He was a truck driver.
And he got tired of waiting for his trucks to be unloaded by hand.
So he invented a box.
A standardised corrugated steel box.
The industry laughed.
“You can’t put everything in a box.”
“Some things are too delicate.”
“Some things are too unique.”
“You’ll always need our expertise.”
They were wrong.
The box didn’t just move goods.
It moved value.
The value used to live in the loading.
The human skill of the port.
The box moved that value into logistics.
Into the system that moved the box.
The shipping container didn’t make dockworkers better at their jobs.
It made their jobs irrelevant.
The ports that survived weren’t the ones with the best loaders.
They were the ones that tore down their docks
and rebuilt their entire geography around the box.
AI is your container.
It is standardising the loading of creativity.
If your agency’s value is “we make things”
you are the dockworker in 1957.
Highly skilled
at a version of shipping
the world no longer needs.
When infrastructure changes,
power moves.
Think about the spreadsheet.
Before Excel, strategy was often dictated
by the loudest voice in the room.
Or the longest tenure.
Or the strongest gut feel.
Then the spreadsheet arrived.
It didn’t just add numbers.
It changed who was allowed to speak.
Suddenly, a 24-year-old analyst could walk into a boardroom,
and dismantle the CEO’s three-year plan.
Not because they were more creative.
But because they had a model.
They could change one variable
and show that the gut feel was a fantasy.
The spreadsheet made judgement visible.
AI is now doing this to your Creative Directors.
For years, the CD has been the high priest of the agency.
Wearing black. Speaking in riddles.
Deciding what’s “good” via an invisible internal compass.
AI is about to expose that process.
It can generate a thousand iterations
based on decades of data
on what wins awards
or what sells soap.
Power is moving away from the creator
and toward the curator.
The question is not whether this happens.
The question is whether your leaders can defend a human judgement
when the machine offers a data-backed alternative.
Or whether you quietly surrender your agency’s soul
to the model.
There is a darker side to infrastructure.
Cognitive offloading.
Consider GPS.
It’s miraculous.
It’s saved billions of hours.
But neurologists have found that heavy GPS users
develop a shrinking hippocampus.
The part of the brain responsible for memory and navigation.
We outsourced direction
and lost the skill of knowing where we are.
We’ve all had that moment.
You’re following the blue line.
And suddenly you realise
you have no idea where you are.
If the phone died,
you’d be lost in your own neighbourhood.
We are doing this to young creative talent.
If a junior uses AI for mood boards, sketches, headlines, ideas
they aren’t learning a tool.
They are atrophying a muscle.
Taste isn’t a gift.
It’s a callus.
It’s formed by friction.
By ninety-nine bad attempts
in service of one good one.
When you remove the struggle,
you remove the learning.
We risk raising creative navigators.
People who can follow a path
but are lost the moment there isn’t one.
The agency of the future must be a gymnasium.
Not just a factory.
You have to protect the struggle.
There is one misconception more dangerous than all the rest.
That leadership in an AI world
is about having the answers.
History says otherwise.
April 13th, 1970.
Apollo 13.
Two hundred thousand miles from Earth,
an oxygen tank explodes.
Alarms.
Power loss.
Rising carbon dioxide.
The mission is over.
The flight director is Gene Kranz.
There is no manual for this.
No precedent.
No best practice.
The computer is useless.
The map is wrong.
And this is where leadership appears.
Kranz doesn’t wait for better data.
He doesn’t ask the computer for suggestions.
He says one thing:
“Failure is not an option.”
Not as bravado.
As responsibility.
What followed wasn’t genius.
It was judgement.
Engineers dumping parts on a table saying,
“This is what they have. Make it work.”
Apollo 13 didn’t succeed because of tools.
It succeeded because someone owned the decision
when the tools failed.
AI puts agencies into Apollo 13 moments every day.
The model will give you an answer.
Often a very confident one.
Leadership begins when you decide
whether to trust it.
Because infrastructure removes friction.
And friction is where responsibility used to live.
With AI, failure becomes ambient.
The system did it.
The model suggested it.
The process allowed it.
That’s the leadership trap.
In an AI organisation, leaders aren’t the best makers.
They’re the ones willing to say:
“This one’s on me.”
AI doesn’t kill leadership.
It exposes it.
When judgement becomes optional,
character becomes visible.
And many leaders have been hiding behind complexity for years.
AI removes the hiding place.
Infrastructure also changes standards.
Before streaming, television was an appointment.
You watched what was on.
Then came infinite choice.
Our patience collapsed.
If something doesn’t grab us in three minutes, we skip.
The infrastructure reshaped storytelling.
Faster.
Louder.
More formulaic.
AI is doing this to your clients.
When variation is infinite,
they stop asking “Is this brilliant?”
and start asking “Is this ready?”
AI pulls us toward the mean.
It is, by design, a machine for average.
If you don’t fight it,
you’ll produce high-speed average.
Efficient.
Polished.
Forgettable.
So where does this leave us?
Infrastructure always does three things.
It moves value.
From making to judging.
It redistributes power.
From priest to curator.
And it exposes leadership.
By the end of 2026, there will be no such thing as an “AI agency.”
It will just be the agency.
The tech will be invisible.
Like electricity in the walls.
The winners won’t be the ones with the most licenses.
They’ll be the ones with the courage to stay human.
Agencies that say:
“We use the box to ship the goods.
But we don’t let the box decide what’s worth shipping.”
In a world of infinite noise,
the only premium left
is human accountability.
AI can give you an answer.
It can’t care if it’s right.
It can’t feel the risk.
It can’t lose its job.
So don’t stare at the waves.
Look at your people.
Look at your culture.
Are you building a dock for hand-loading crates?
Or a new creative architecture?
One that frees humans
to do the only thing machines can’t.
To be brave.
To be weird.
And to be right.
The tide is coming in.
It’s time to start building.
Thank you.





