What governance can learn from AI, Mexico and the messy reality of delivery
Sean, a creative strategist with an MBA and a talent for spotting what others overlook, offered a quiet but powerful insight during a Project Planning Governance meeting in June 2025:
“It’s not ‘Did we do it?’ but ‘How did we do it?’ that matters.”
At first glance, it could sound like another phrase destined for a workshop wall.
It was not.
It was a gentle rebuke to a culture that can too easily reduce delivery to completion. Task done. Box ticked. Paper approved. Milestone met. Move on.
But Sean’s point asked for something deeper. It asked us to look at the ethics of delivery, not just the evidence of delivery. It asked what gets hidden when we celebrate speed, control and completion without asking what those things cost.
Because governance is not only there to confirm that something happened.
It should help us understand what the way it happened tells us.
Delivery is not a tick box
Much like AI systems that confidently complete a prompt with something plausible but wrong, organisational delivery can produce outputs that look right, feel right and still miss the point.
“Done” is not the end of the story.
Sometimes it is where the real scrutiny should begin.
Inside the ticked box may sit unseen debts: to people who were excluded, to timeframes that were compressed, to values that were quietly compromised, to risks that were managed by passing pressure somewhere else in the system.
We tend to celebrate outputs.
Reports delivered.
Deadlines met.
Buildings opened.
Systems implemented.
Savings reported.
Boards reassured.
But the harder question is often:
“What changed in order for us to say this was done?”
Who was left out of the room?
What was rushed?
What was not understood?
What harm was accepted because it was hard to measure?
What concern was softened because the project needed to keep moving?
These are not anti-delivery questions.
They are better delivery questions.
And they matter even more as AI becomes part of organisational work. When AI mirrors our methods, it does not only reflect our goals. It reflects our blind spots. If we teach machines to optimise for completion without teaching them to ask about method, fairness and consequence, we should not be surprised when they reproduce the shallowest version of our governance.
If AI learns from us, as I have argued elsewhere in pieces such as The Lighthouse Fallacy and Follow the Noise, then what are we teaching it when we reward projects simply for looking complete?
Are we training systems to value appearance over substance?
Are we building tools that will help us ask better questions, or tools that will help us produce more confident answers to the wrong ones?
The answer lies in governance that is not only planned, but reflected upon, challenged and open to difficult truths.
Governance, at its best, becomes an organisational conscience.
A quiet voice asking:
“Just because we could, should we have?”
And perhaps more importantly:
“Did we listen when someone said, ‘I’m not sure this is right’?”
When the plan is still moving
I was reminded of this during a detailed discussion about how we had approached one of our more complex projects.
The account was full of commendable structure. Stakeholders had been coordinated. Changes had been documented. Test runs had been planned. Outcomes had been tracked. The governance artefacts were there.
It ticked the boxes.
But that was not the whole truth.
The real achievement was not that the plan unfolded smoothly. It was that the plan had to be implemented and reimplemented while the ground kept moving.
Technical issues.
Deadline pressure.
Stakeholder turnover.
Unexpected dependencies.
A system that kept shifting as we tried to stabilise it.
It was, to borrow an imperfect but useful analogy, less like executing a clean project plan and more like the opening sequence of Saving Private Ryan.
The plan exists. The objective is clear. The troops know where they are meant to go.
Then the ramp drops.
Bullets tear through the surf. Soldiers spill into the sea, dragged down by their packs. Orders disappear into noise. Explosions turn sand into shrapnel. The camera does not present strategy as a tidy diagram. It shows the moment where planning meets violence, confusion and survival.
This is not a plan unfolding.
It is a plan meeting reality.
That is where Sean’s point matters.
If we only ask, “Did we do it?”, we miss the intelligence of the delivery. We miss the improvisation, the judgement, the recovery, the relationships and the small decisions that kept the work alive.
Helmuth von Moltke the Elder is often paraphrased as saying that no plan survives contact with the enemy. In delivery, no plan survives contact with the organisation unchanged.
That does not mean planning failed.
It means planning was only the opening move.
The deeper success may be that the plan bent without breaking. It adapted in real time. People made judgements under pressure. They held enough shared purpose to keep moving when the original route was no longer enough.
That is the “how” governance should care about.
Mexico, maps and method
Consider the aqueduct of Padre Tembleque, a 16th-century engineering achievement in central Mexico.
It was not only that the aqueduct was built.
It was how.
Through collaboration between Indigenous communities and Franciscan friars. Through practical experimentation. Through local knowledge, shared labour and engineering ambition. Through a process that was technical, social and spiritual at once.
The method mattered.
The structure still stands, but the deeper lesson is not just the stonework. It is the form of collaboration that made the stonework possible.
There was no project management system. No dashboard. No automated milestone tracker.
But there was governance of a kind: relationship, purpose, accountability, negotiation and shared commitment to an outcome that mattered.
We like to think our modern systems are more sophisticated.
In some ways, they are.
But sophistication can become shallow if it forgets relationship. A perfectly tracked project can still fail people. A beautifully formatted plan can still exclude the voices that would have made it better. A completed output can still leave damage behind it.
If you walk the length of the aqueduct today, you see more than infrastructure.
You see a story.
A story of alignment across languages, worldviews, technologies and needs. A story that endured because the people involved did not only care about the end. They cared about the means.
Modern infrastructure, digital or physical, needs the same honouring of the “how”.
Not because it slows delivery down.
Because it makes delivery more trustworthy.
People may forgive honest delay.
They struggle to forgive hidden harm dressed up as progress.
They remember being bypassed. They remember being unheard. They remember consultation that was really a performance. They remember when the method told them they did not matter.
The performance of success
Twentieth-century cinema understood this problem well.
In Chaplin’s Modern Times, we see a man trapped in the cogs of an industrial machine, tightening bolts faster and faster until meaning disappears. The task is being performed. The work is being done. The machinery is operating.
But at what cost?
His agency, his dignity, his ability to question the system around him.
The comedy works because the logic is recognisable. The process has consumed the person.
Or think of The Truman Show. Every illusion is beautifully executed. The production is scheduled, budgeted, staged and broadcast. From a management perspective, the operation is extraordinary.
But ethically, it is hollow.
The project is a success until someone asks the wrong question.
Or perhaps the right one.
That is the point. Good governance, like good storytelling, dares to ask the question that disrupts the performance.
Appearance is not substance.
Control is not care.
Certainty is not wisdom.
A completed project is not automatically a good project.
Governance should listen for the dissonant note in the chorus of agreement. It should ask whether the people at the edge of the frame are seeing something the centre has missed.
From AI to organisational intelligence
As councils and other public bodies experiment with algorithmic tools and automation, we must ask not only whether they work, but how they work.
Are the algorithms trained on unjust patterns?
Are the outcomes explainable?
Are the metrics meaningful, or merely convenient?
Do the tools support judgement, or quietly replace it?
Do they help us lead, or do they simply help us justify decisions already made?
Too often, governance arrives too late.
It turns up like a forensic team at a project post-mortem. It asks what went wrong after the damage has already happened.
What we need is more governance as pre-mortem.
Speculative. Curious. Preventative.
What might go wrong?
Who might be harmed?
Who is not in the room?
What data are we missing?
What assumption are we treating as fact?
What silence is trying to tell us something?
Good planning governance does not simply sign off solutions.
It creates the conditions for intelligent challenge.
It builds coalitions of doubt and curiosity. It asks questions that make us uncomfortable in useful ways. As I explored in Thinking While Falling and Nobody Died, leadership requires us to treat uncertainty not as a defect to be edited out, but as information we need to work with.
This is also where AI can help us, paradoxically.
Not because it knows better.
But because it can hold a mirror to our patterns.
If we train it only on outcomes, we will get systems that value completion. If we train it on thoughtful process, challenge and reflection, we may get tools that help us notice what we usually miss.
We should not ask AI only:
“What did we do?”
We should ask:
“How did we do it, and what did that method reveal?”
That may be where organisational intelligence begins.
How, not just if
Planning governance should not be the gate at the end of the journey.
It should be the table we keep returning to.
Not only to ask:
“Did it work?”
But:
“How did we get here?”
Who travelled with us?
Who was left behind?
What changed in the process?
What did we learn?
What would we do differently?
What did our method reveal about our values?
Governance must become less like a stamp of approval and more like a forum for reflective storytelling.
Because the way we deliver matters.
Process, as much as product, defines who we become.
That does not mean delivery should become indulgent, slow or endlessly self-examining. Public services need action. Residents need outcomes. Organisations need decisions.
But outcomes without reflection can become hollow.
And speed without method can become damage wearing the mask of progress.
The lesson taught by Chaplin’s bolt-tightening, Padre Tembleque’s aqueduct and the messy reality of delivery is the same lesson we should teach our machines too:
The how matters.
Not as decoration.
Not as a lesson learned afterthought.
As the thing that tells us whether the delivery was worthy of the outcome.
Sean’s insight deserves to stay with us because it refuses the easy comfort of completion.
It reminds us that governance is not only about whether the work crossed the line.
It is about what kind of organisation crossed it.
And what it became on the way.
Comments
Post a Comment