Nobody made the decision.
At least, nobody can quite tell you who did.
The threshold changed. The service shifted. The application was rejected. The workflow now behaves differently.
Ask around and you’ll hear the same strangely modern phrases:
- “That’s what the system says.”
- “It came through the platform.”
- “The model scored it that way.”
- “We can’t progress it until the dashboard updates.”
No one is fully responsible, and yet everyone is affected.
I find this one of the most unsettling features of modern organisational life.
Power hasn’t disappeared. If anything, it has become stronger.
It has just become harder to point at.
It sits inside systems.
Inside platforms.
Inside dropdown menus and workflow rules.
Inside decisions made six months ago during configuration workshops that now feel like natural law.
If the earlier lenses showed us organisations as machines, organisms, cultures and political systems, this is where the modern picture becomes distinctly stranger.
Because increasingly, organisations behave like algorithms.
We Don’t Just Use Systems. We Live Inside Them.
Gareth Morgan gave us extraordinary metaphors in Images of Organization, but he was writing before the world we now take for granted.
He could see bureaucracy. He could see politics. He could see domination and culture.
He could not fully see a workplace where a workflow engine might shape behaviour more effectively than a senior manager.
But that is where we are.
A great deal of management now happens through systems that quietly decide:
- who gets prioritised
- what gets escalated
- how performance is measured
- what counts as success
- which exceptions are allowed
- what becomes invisible
Sometimes these are labelled AI.
Often they are not.
A spreadsheet can be an algorithm.
A procurement route can be an algorithm.
A housing threshold can be an algorithm.
A dashboard can govern behaviour more effectively than a policy ever could.
The important thing is not the sophistication of the technology.
It is the logic embedded inside it.
We Pretend the System Is Neutral
One of the most dangerous phrases in governance is:
“The system won’t allow it.”
It sounds objective. Final. Almost moral.
As though the system has arrived from the mountain carrying stone tablets.
But systems are never neutral.
Someone chose the threshold.
Someone designed the approval route.
Someone decided which fields were mandatory and which outcomes were measured.
Power has not vanished.
It has been hidden behind interface design.
That matters because hidden power is much harder to challenge than visible authority.
We know how to disagree with a director.
How do we disagree with a dropdown menu?
This is why I’ve always been suspicious of organisational certainty dressed as technical necessity.
Sometimes “best practice” is just an old political decision with better branding.
The Dashboard as Modern Theology
I sometimes think dashboards have become the new religion of management.
We gather around them looking for revelation.
The red box means danger.
The green box means reassurance.
The arrow points upward, therefore progress exists.
I say this as someone who writes board reporting and likes good evidence.
Metrics matter.
But metrics are stories disguised as numbers.
They tell us what somebody decided was worth counting.
They also tell us what has been made invisible.
If throughput becomes the dominant measure, quality disappears.
If adoption becomes logins, understanding disappears.
If speed becomes success, reflection starts to look like failure.
We begin managing the representation of reality rather than reality itself.
And that is when the algorithm starts managing us.
I wrote something similar in The Fiction of Control. Contracts, dashboards and governance playbooks often promise certainty, but what they really reveal is how much we are trying to manage ambiguity by turning it into measurable ritual.
Sometimes the ritual becomes the work.
AI Doesn’t Create This. It Reveals It.
This connects directly to what I’ve been writing about AI.
I keep returning to the same argument: AI is not suddenly changing organisational life. It is exposing how it already worked.
We have always embedded values inside systems.
AI simply makes the process faster, larger and harder to ignore.
A hiring process that quietly rewarded sameness can now automate it.
A service threshold that informally disadvantaged certain groups can now operationalise that disadvantage at scale.
The algorithm does not invent the politics.
It industrialises it.
This is why conversations about AI ethics that focus only on the model miss the point.
The real question is rarely “is the AI biased?”
It is:
What assumptions were already sitting in the organisation waiting to be accelerated?
That is a much less comfortable conversation.
Which is usually a sign it is the right one.
The Prime Directive Problem
This also links back to my old obsession with the Prime Directive in Star Trek.
The interesting question was never whether intervention was possible.
It was who had the right to decide what counted as ethical intervention.
AI gives us the same problem.
Who decides what fairness looks like?
Who decides what risk is acceptable?
Who decides whether efficiency should outrank dignity?
Those are not technical questions.
They are moral and political ones disguised as implementation choices.
And because they are hidden inside systems, they often escape the scrutiny they deserve.
Local Government Is Full of This
In local authority work, this is everywhere.
Permit schemes.
Housing allocation.
Adult social care prioritisation.
Translation services.
Procurement thresholds.
Contract management triggers.
These are not neutral workflows.
They are policy decisions wearing technical clothes.
If we pretend otherwise, we create the perfect conditions for unfairness with plausible deniability.
Nobody intended harm.
The system simply behaved as designed.
That sentence should make all of us uncomfortable.
The Quiet Loss of Ownership
The other danger is responsibility dissolving.
The supplier built it.
IT configured it.
The business approved it.
Procurement bought it.
Operations use it.
Everyone touched the decision.
Nobody feels they own it.
This is where governance has to grow up.
Accountability cannot be outsourced to software.
A poor automated decision is still our decision.
We just made it earlier, during design.
Usually in a meeting nobody remembers.
Usually under the heading “configuration.”
That should worry us more than it does.
Prompting, Photography and Editing
I’ve been thinking recently that prompting AI is like photography.
You choose the frame. You decide what enters the picture.
But editing—that’s where meaning is shaped.
The same is true here.
The system is the photograph.
The defaults, thresholds and reporting logic are the edit.
They determine what becomes visible, what gets cropped out and what the audience is encouraged to believe matters most.
Which means governance is not just about approval.
It is about authorship.
Who framed this?
Who edited it?
Who decided what we would be allowed to see?
That is not a technical question.
It is leadership.
A Different Question Again
So the question evolves.
Not:
“Who made this decision?”
But:
What assumptions have we embedded so deeply that they now look like facts?
That question is harder because it forces us to challenge invisible architecture.
It means governance cannot stop at stage gates and board papers.
Sometimes governance means interrogating a mandatory field.
Sometimes it means asking why the default setting exists.
Sometimes it means refusing to accept “the system says” as the end of the conversation.
And frankly, that may be one of the most important leadership habits we can build.
Where This Leaves Us
The machine gave us structure.
The organism gave us adaptation.
Culture gave us meaning.
Politics gave us power.
The algorithm gives us something more unsettling.
Power without a face.
Rules without a visible author.
Decisions that feel inevitable because we forgot they were designed.
And perhaps that is where modern leadership becomes hardest.
Not managing people.
Remembering that the systems we build are arguments about how the world should work.
Part VI: When the Narrative takes the Chair
Comments
Post a Comment