Governed Is Not Controlled.
Written and narrated by Serguei Poppeleer.
There was a time when you could set a project up once and trust the setup to hold. You scoped it, contracted it, governed it, and executed against a world that mostly stood still while you worked.
That world is gone.
A tariff appears overnight. A currency moves. A supplier’s government rewrites its export rules. A shipping route closes. A sanction lands on a counterparty you signed with last quarter. The assumptions your project was built on start expiring mid-flight — and no amount of good governance written at kickoff can see them coming.
The annual strategy review used to be enough. Assess the landscape once, set direction, execute, and handle the rare shock when it came. Now the landscape changes between board meetings. The organisations that wait for the next review to react are already behind the ones reading the market as it moves.
And here is the part nobody says out loud: most projects were never built to absorb that. They were built to be governed — reported on, reviewed, signed off.
This report is about that difference — and about a simple structure that lets you hold control no matter where a project sits, or how fast the ground moves under it.
The checkpoint illusion
The belief that quietly runs most capital programmes is this: get the setup right at the start, govern it through stage gates, and delivery will follow. It is a comforting belief. It is also wrong in exactly the conditions we now work in.
Control gets treated as a checkpoint you pass.
Stage gate cleared, RAID log updated, steering deck submitted — control confirmed, move on. But a checkpoint only tells you the state of the project at the moment you looked. It says nothing about the next four weeks, when the supplier’s lead time doubles, the host government revises a rule, or the assumption your business case rested on stops being true.
A project does not fail at the checkpoint. It fails in the space between checkpoints — while everyone is still looking at the last green light.
Cross-border makes it worse. The further your project reaches across borders, jurisdictions, currencies and cultures, the more of those gaps exist, and the faster reality diverges from the plan. Frontier markets do not create this problem. They just expose it faster. And as the world fragments, more markets start behaving like frontier markets — so the exposure spreads to places that used to feel safe.
The checkpoint model was built for a stable, globalised world. We do not live there anymore.
One structure. Four entry points. Five rails.
Every capital project or operation is in one of four states. Each state needs a different intervention — but all four run on the same underlying mechanics. We call the structure the Control Ladder.
You are taking an idea across the line into a funded, contracted, operating project, often in a market you have not run before. The risk: building on assumptions nobody wrote down, in conditions you have not tested.
Delays, claims, scope arguments, vendor slippage. The risk: losing time and credibility while everyone debates what was actually agreed.
The world moved — new tariffs, new route, new partner, new rules — and the setup no longer fits. The risk: executing perfectly against a plan that is now pointed at the wrong target.
Scale without losing control. The risk: growth outrunning the system, so volume breaks the very controls that made you reliable.
The 5Cs — the rails that run the full height
The 5Cs Framework is the same whether you are starting, resolving, reshaping or growing. It is what makes the ladder a ladder, and not four separate services bolted together.
One written scope — assumptions, acceptance criteria, explicit out-of-scope — locked before work starts.
One operating plan that matches site and production reality week to week. Not a deck that diverges by week two.
Every open item gets an owner, a decision-by date and a stated consequence. Decisions close before they cascade.
Commitments verified against capacity, lead times and readiness — not accepted on a promise.
A fixed rhythm — same inputs, same meeting, same outputs. Slippage caught while there is still room to move.
Why one assessment a year is no longer enough
The old operating assumption was that context is stable: set the controls once, run them, and review strategy annually. That assumption is dead. Context now changes faster than any review cycle can catch it. Supply chains are regionalising. Jurisdictions are diverging. Currencies and trade rules move on political time, not project time.
So the question is no longer only whether you are on track against the plan. It is whether the plan is still pointed at the right target — and that has to be asked continuously, not once a year.
The best organisations no longer assess and then act. They react as the market unfolds — because the cost of waiting for the next review is now higher than the cost of constant attention.
The Control Ladder is built for exactly this. When the world moves, you do not start a new project. You move across a rung — RESOLVE the immediate exposure, then RESHAPE the setup to the new reality — on the same rails, without losing the thread.
Where AI fits — and where it does not
Continuous control has always had one problem: it is expensive. You cannot put a person on every assumption, every lead time, every rule change, re-checking the whole picture every week. That is why, historically, we settled for the annual review. It was all we could afford.
AI changes the economics. It is what makes continuous control affordable for the first time. In what we build for clients, AI does the watching that humans cannot sustain — monitoring the inputs that feed the 5Cs, flagging when a vendor’s signal drifts from the confirmed date, when a market input moves against a business-case assumption, when a decision has sat open past its date. It surfaces the issue while it is still small, so a human can make the call early instead of explaining the fire late.
What AI does not do is replace the judgement. Automate a broken process and you get faster chaos, not control. You standardise the inputs first — the 5Cs — and only then let AI accelerate them. The sequence is not optional: architecture before automation, every time.
AI does not run the control system. It enforces it — faster than a person can, and without getting tired.
The commercial layer — control you can hold someone to
A control system that lives only in a project plan is a suggestion. To be real, it has to be commercial — written into the terms, conditions and KPIs that govern money and obligation. This is where the operational set-up meets the contract. The operational system we build to produce what a client expects is managed, commercially, through the same five rails — translated into bespoke terms, not off-the-shelf clauses, because every cross-border project has its own fault lines.
Scope, assumptions and out-of-scope set as the contractual baseline, with a defined change mechanism — so scope cannot move quietly.
The operating plan named as the contractual baseline programme everyone is measured against.
Decision rights, decision-by dates and consequences embedded in the governance terms — so “we’ll align later” carries a cost.
Supplier commitments tied to verifiable readiness criteria and lead-time obligations, with remedies that bite.
KPIs and reporting cadence written in; leading indicators set as triggers — so the contract reacts to drift, not only to failure.
Operations produce the result. The contract makes the result enforceable. The 5Cs are what connect the two.
What to do with this
Be honest about which of the four states each major project is actually in — START, RESOLVE, RESHAPE or GROW. Most portfolios have projects on at least two rungs at once.
For each project, ask the five questions: scope contained, one plan running, decisions closing, vendor dates confirmed, cadence fixed? That tells you where control is real and where it is theatre.
Replace the once-a-year review with a standing rhythm that reads the market continuously — and decide, in advance, which signals force a RESHAPE.
List the assumptions your business case depends on, and the market conditions that would break each one. This is what AI then monitors on your behalf.
Standardise the 5C inputs before automating anything. Architecture first — otherwise you are buying faster chaos at a premium.
Translate the 5Cs into the contract: baseline scope, baseline plan, decision terms, verified vendor obligations, KPI triggers. If it is not in the terms, it is not enforceable.
Assume divergence, not convergence. The further a project reaches, the more the rails have to do — so design them to carry the load before you need them to.
The objections worth answering
“This sounds like more process.”
It is less. The 5Cs replace governance theatre — the packs nobody reads — with five mechanisms that do real work. You keep your stack; you make it enforceable.
“We already have governance.”
Everyone does. Every project that failed had governance, and most had a good one. The question is not whether you have it. It is whether it holds when the ground moves.
“AI will take the judgement out of it.”
The opposite. AI takes the watching out of it — the unsustainable, around-the-clock monitoring — so your people spend their judgement on decisions, not on hunting for the problem.
“Our markets are stable.”
For now. The point of the ladder is that you do not have to predict which shock arrives. You build a system that can move rungs when one does.
“We can’t rewrite our contracts mid-flight.”
You do not rewrite everything. You target the five points where control actually breaks and make those enforceable. Usually that is a handful of terms, not a new contract.
Proof from the field
Remote terrain, frontier market, no established supply chain. Every standard assumption pointed to a twelve-month timeline at significant cost, with a high probability of claims. What happened instead: with scope contained, one plan shared, decisions forced to close and commitments verified, the camp was built in roughly six months at a fraction of the typical cost. No claims were filed. The local community built the fence. The state client released payment on a single note — “We have started the works.”
The point is not the geography. It is what frontier conditions expose: when the rails hold, delivery becomes predictable even where most projects drift.
Commercial controls had to hold across multiple jurisdictions, hostile contracting environments and complex multi-party interfaces. The 5Cs, written into the commercial structure rather than left in a plan, were what kept the programme defensible when the environment pushed back.
None of these worked because the market was easy. They worked because the system was built to hold when it wasn’t.
Find your rung.
The CONTROL diagnostic maps you across the four entry points — START, RESOLVE, RESHAPE, GROW — and shows you where control is real and where it is exposed. Twenty questions. Four minutes. Instant result.
Start with the CONTROL diagnostic → Enter the library →