Sometimes the organisation works exactly as designed.
That is the problem.
- The targets are being met.
- The dashboard is green.
- The contract is compliant.
- The governance is defensible.
And yet, if we are honest, something feels wrong.
- People are exhausted.
- Frontline teams are quietly carrying decisions made somewhere else.
- Communities feel the service has become harder to reach, not easier.
- The system is succeeding in a way that makes success feel morally awkward.
This is one of the more uncomfortable lessons in organisational life.
Failure is not always the danger.
Sometimes the greater risk is a system that performs efficiently while producing outcomes we would struggle to defend if we looked at them directly.
If the machine gave us structure, the organism adaptation, culture meaning, politics power, the algorithm hidden authority, narrative shared reality and the network interdependence, this final lens asks a harder question.
Not “does it work?”
But:
Who is it working for?
Because sometimes organisations are not simply systems of coordination.
They are instruments of domination.
That word sounds dramatic. It is supposed to.
We Prefer Kinder Words
Gareth Morgan deliberately used this metaphor in Images of Organization because polite language can hide hard truths.
He argued that organisations can function as instruments of domination: systems that concentrate benefit for some while externalising cost onto others.
We tend not to say it like that.
We prefer gentler phrases.
- Efficiency.
- Transformation.
- Channel shift.
- Resilience.
- Optimisation.
These are not false words.
But they are often incomplete ones.
We rarely say:
“This decision makes life harder for people with the least room to absorb it.”
We say:
We are driving efficiencies.
We rarely say:
“This policy assumes everyone has equal capacity to navigate bureaucracy.”
We say:
We are improving customer journeys.
We rarely say:
“This governance decision protects senior assurance by moving operational pain downward.”
We say:
We are strengthening control.
Language matters because incomplete language allows domination to become administratively respectable.
That is how systems stay clean on paper and messy in real life.
Complexity Always Travels Downhill
There is a pattern to this.
Complexity travels downward.
The strategic decision sits at the top.
The practical consequence lands at the front line.
A policy sounds elegant in committee.
On Monday morning, it becomes three extra steps for the person answering the phone.
A procurement route protects governance and auditability.
It also creates months of friction for the operational team trying to deliver an actual service.
An AI-enabled efficiency improves reporting.
It quietly removes discretion from the person who understands the human context best.
From above, the system looks cleaner.
Because the mess has been relocated below.
Usually to someone with less authority to challenge it.
That is not always malicious.
Often it is simply invisible.
Which can be worse.
Because invisible costs are easier to defend.
I think this sits underneath a lot of what I wrote in The safest decision is rarely the best argued one.
Safe for whom?
That is the question we often avoid.
We Call It Resilience When We Mean Endurance
There is a phrase I have become suspicious of.
“Teams need to be more resilient.”
Sometimes they do.
But often resilience is a polite managerial way of saying:
“We have decided the system will not change, so you must absorb it.”
Frontline staff become experts at carrying complexity they did not create.
Residents become experts at navigating services designed around institutional convenience rather than human experience.
People are praised for coping with avoidable friction.
That is not resilience.
That is endurance.
And endurance has a moral cost.
The danger is that good people become the shock absorbers for badly designed systems.
The organisation then mistakes their professionalism for proof that the design works.
It doesn’t.
It means someone is paying quietly.
Mexico Taught Me This Before Management Theory Did
Some of the clearest lessons about power never came from management books.
They came from Mexico.
Spend enough time there and you notice how beautifully inequality can be normalised.
- Who waits.
- Who explains.
- Who serves.
- Who is expected to be grateful.
It is rarely announced.
It is embedded.
The system feels natural because it has been narrated that way for so long.
Organisations are similar.
Hierarchy often survives not because people defend it loudly, but because it becomes ordinary enough to disappear from view.
That is why I keep coming back to narrative.
Who gets to describe the system determines whether the system looks fair.
Chimamanda Ngozi Adichie warned about the danger of the single story.
One version of reality becomes so dominant that alternatives sound unreasonable.
That is how inequality protects itself.
Very politely.
Modern Times Was Never About the Machine
I said I would not keep using Modern Times, but it refuses to leave.
People remember the conveyor belt.
The real point is alienation.
The machine is not frightening because it is mechanical.
It is frightening because people become extensions of its logic.
That has not gone away.
The conveyor belt is now a workflow system, a performance dashboard, an automated threshold, an AI triage model.
The question remains the same:
Are people using the system, or is the system using people?
That question matters enormously in public service.
Because once people become units of throughput, governance may still be excellent while justice quietly leaves the room.
AI and the Industrialisation of Old Inequalities
This is where AI conversations often become strangely polite.
We talk about productivity and innovation.
Necessary conversations.
But not sufficient ones.
Who benefits from the productivity gain?
Who loses discretion?
Who gets monitored more closely?
Who becomes more replaceable?
The algorithm does not arrive in a neutral world.
It lands inside existing inequalities and usually follows gravity.
Power accumulates upward. Friction travels downward.
This is why DEI cannot be treated as moral decoration added after the design.
It is governance.
If we are not asking who carries the hidden cost of automation, we are not doing governance.
We are doing theatre with a dashboard.
That links directly to my article on AI and the Prime Directive.
The ethical question was never simply “can we intervene?”
It was “who gets to define what good looks like?”
The same applies here.
- Efficiency for whom?
- Fairness by whose measure?
Those are leadership questions, not technical ones.
Our Own Work Quietly Leads Here
Even our own Impact → Interpretation → Risk work pushes us toward this.
Impact is never neutral.
Change lands differently depending on where you stand.
A “small improvement” for a sponsor may be a daily frustration for frontline staff.
A “low-risk transition” in the board paper may be deeply destabilising for the people whose work is being redesigned.
Interpretation matters because position matters.
Risk is often just inequality described formally.
That should make us much more cautious when we use the word proportionate.
Proportionate for whom?
That is the adult question.
A Different Leadership Test
Leadership here is not simply delivery.
It is moral attention.
Can we see the burden our own governance hides?
Can we notice when a system is functioning perfectly and still producing the wrong outcome?
Can we ask who is paying for the certainty we are presenting upward?
That is harder than managing milestones.
Because it requires us to question systems that are often rewarding us for maintaining them.
Which is why it is real leadership.
Not performance.
Judgement.
Not “did we deliver?”
But “should this have been delivered like this?”
That is a much less comfortable board question.
Which is probably why it is the right one.
The Final Question
So perhaps the final question in this whole series is the simplest and the most uncomfortable.
Not:
“Is the programme succeeding?”
But:
Who is carrying the cost of that success?
Because every organisation answers that question whether it intends to or not.
And the answer tells us far more about its values than any strategy document ever will.
The machine taught us structure.
The organism taught us adaptation.
Culture taught us meaning.
Politics taught us conflict.
The algorithm taught us hidden power.
Narrative taught us shared reality.
The network taught us interdependence.
Domination forces us to look at consequence.
- Who benefits.
- Who absorbs the risk.
- Who gets to call the outcome a success.
That may be the least comfortable metaphor of all.
Which is probably why it matters most.
Part IX: The non-appearance of the Psychic Prison
Comments
Post a Comment