The pattern that kept bothering me
I have spent much of my career helping organisations change. Over the years, one pattern has bothered me more and more. Capable leadership teams spend enormous amounts of time, money and energy fixing problems that never really go away.
Innovation is too incremental, so they run an innovation programme. Decisions are too slow, so they clarify decision rights and set up a RACI. People are not taking ownership, so they launch an empowerment initiative. Functions are not collaborating, so they change the structure.
Often the diagnosis sounds reasonable, and the intervention does too. And yet two or three years later the same frustration is back, sometimes in the same form, sometimes wearing different clothes.
At some point I started asking the uncomfortable question.
Are we treating the cause, or putting another plaster on the symptom?
That question is what led to the Adaptability Audit. What follows is the thinking behind it. First, what adaptability actually is, and the precise point where it tends to break. Then why that point is so hard to find, both in most organisations and increasingly in ones where people and AI systems now work side by side. And finally how the Audit locates it, and traces the cause and the loop that keep it in place. The outcome is a short, decision-ready view of what to change first, what to protect, and what to transfer elsewhere.
What adaptability actually means
Adaptability is the capacity to keep an organisation fit with changing reality, by converting relevant signals into appropriately timed changes in commitments, operating behaviour and, when necessary, the organisational system itself.
Fit with changing reality means adaptability is contextual. Sometimes adaptability calls for experimentation and speed, sometimes for reliability, discipline and control. A nuclear power station should not behave like an early-stage technology company.
Relevant signals means noticing what matters before it becomes obvious. More data does not automatically help. Data describes what has already happened, and forecasts model what might. Neither guarantees that anyone notices a shifting customer need or an assumption quietly going out of date.
Appropriately timed means faster is not always better. Sometimes the right response is immediate. Sometimes it is to learn, wait, or protect stability while the picture clears.
And there are three depths of response. Commitments are the choices about priorities and resources: what to fund, pursue or stop. Operating behaviour is what people actually do differently in the work: how decisions get made, how customers get served, how resources move. The organisational system is deeper still: structures, decision rights, incentives, measures, processes, technology, governance and the resulting unwritten rules that shape behaviour. Sometimes adapting means changing the system that keeps recreating yesterday’s behaviour.
Put simply, adaptability is a chain:
Notice what matters, turn it into the right commitments, change how the work is actually done, and when necessary redesign the system itself.
That chain is the key to everything that follows, because it can break at any link. The breakpoint is the precise point where the chain fails: where reality stops becoming action.
Why this is hard to diagnose
If the breakpoint were easy to see, none of this would be necessary. It is hard for three reasons.
First, the real cause usually sits deep. A slow decision is rarely just a slow decision. Underneath it sit decision rights, incentives, budgeting rules, the way the organisation senses its market, the unwritten rules people actually follow, the underlying assumptions leaders hold without noticing, and the loops that quietly keep all of it in place.
Behaviour is evidence. It is not yet a diagnosis.
A conversation I had recently with someone in the Dutch government brought this home. He described how easily organisations say that civil servants need more lef, more courage to stick their necks out, challenge convention and take initiative.
But “people need more courage” is already a diagnosis. It assumes the behaviour itself is the problem. A better starting point is to ask what makes holding back rational. What happens when someone challenges a decision? What happens after a reasonable mistake? Where does authority actually sit? Which targets matter when priorities conflict? What have people learned from previous attempts to change things?
The same applies to familiar diagnoses such as “we have silos”, “leaders protect their own turf”, “people don’t collaborate” or “decision-making is too slow”. These may accurately describe what we see, but they do not yet explain why it happens.
Across several episodes, a very different picture may emerge. What initially looked like poor collaboration might turn out to be a system in which functional targets, performance reviews and resource-allocation mechanisms make local optimisation perfectly rational whenever enterprise priorities require trade-offs.
That distinction matters because the intervention changes completely. If the problem is a lack of courage, we might invest in leadership training or encourage people to speak up. If the system repeatedly punishes the very behaviour we are asking for, more exhortation will achieve little.
This is why diagnosis should begin with real episodes and work backwards from behaviour to the mechanisms producing it. Finding a cause at that depth draws on many different bodies of knowledge: disruption theory, psychological safety, dilemma thinking, and how good people get trapped in bad loops. No single framework holds them all.
Second, the system itself is changing. Organisations are becoming hybrids of people, AI tools and autonomous agents. A signal may be filtered or summarised by a machine before anyone sees it. A recommendation engine may quietly optimise for last year’s objective. An AI-enabled workflow may still sit inside old approvals and old KPIs. The root cause is no longer necessarily in human behaviour alone.
Third, the usual instrument is a survey, and surveys struggle at both ends. On the way in, people are survey-tired, and a broad statement like “decisions are made quickly here” tells you very little about why one particular decision stalled. On the way out, a dashboard of scores and a laundry list of improvement actions still leaves management with the hardest question unanswered: what do we change first?
This is the gap the Adaptability Audit is built to close.
Unlike a culture survey, which hands you a dozen things to improve, the Audit finds the breakpoint, and the one or two changes that make the real difference.
How the Adaptability Audit is different
The Audit does not stop at naming the problem, or even its underlying mechanism. It looks for the smallest change capable of shifting the system that keeps reproducing the problem.
It follows one chain:
Strategy and required performance → breakpoint and condition → root cause → leverage point → action.
Five moves run along that chain:
- Start with the performance the strategy actually requires.
- Find the breakpoint, and the organisational condition behind it.
- Trace the root cause beneath it.
- Find the leverage point.
- Decide what to change, what to protect, and what to transfer.
By the end, a leadership team can see where performance is constrained, why capable people behave as they do, which parts of the management system produce that behaviour, and the small number of changes most likely to release it.
How the analysis works
Real episodes, not Likert scales, are the engine. Instead of moving everyone through the same fixed questionnaire, the Audit reconstructs what actually happened in a real situation and lets each answer decide where to probe next.
The Audit deliberately captures what the person initially thinks is causing the problem, and often what they already think should be done about it. That initial explanation is evidence to investigate, not the diagnosis. The conversation then pressure-tests it against what actually happened. Across several episodes, the pattern can look very different from the original explanation.
What people initially say → what the evidence may reveal
>
“People need more courage.” → Authority is unsafe to use.
>
“We have silos.” → Targets reward local optimisation.
>
“Leaders protect their turf.” → Resources remain structurally tied to the core.
>
“We need faster decisions.” → Approval rules make escalation rational.
The Audit therefore moves from what people think the problem is to what the evidence suggests is producing it. One episode can reveal a mechanism. Several episodes show whether it is a recurring system pattern, an exception, or simply the first plausible explanation.
1. Start with strategy and required performance
The same practice can be a strength in one context and a constraint in another. Decentralised decisions add speed near the customer but create real risk where decisions are interdependent and hard to reverse. So the first question is about fit: does this practice serve the performance the strategy now requires? This matters because yesterday’s strength can become tomorrow’s constraint. A system built brilliantly for scale and predictability can become exactly what slows an organisation when its strategy starts to require experimentation or renewal. Starting here also guards against a common error, assuming every performance problem is organisational when the real constraint is capability, capital, regulation or market structure.
2. Find the breakpoint and the condition behind it
Culture surveys tell organisations to become more innovative, collaborative or customer-centric. Those are desired outcomes, and they say little about what is preventing them. The more useful question is whether the organisation has the conditions to produce the performance it needs. The Audit looks across six.
The six organisational conditions, their core failure and what it feels like from the inside:
- ENERGY & Contribution — Will people bring their full contribution? The organisation leaves human capability untapped. “I could contribute more, but I don’t.”
- TRUTH & Sensing — Can we see? The organisation loses contact with reality. “We are acting on the wrong understanding of reality.”
- AGENCY & Empowerment — Can they act? Power sits somewhere else. “I want to act on what I know, but I can’t.”
- CHOICE & Resource Allocation — Can we choose? The company's machinery for making choices distorts the choices. “We understand reality, but our machinery produces the wrong decision.”
- EMBEDDING & Execution — Can we make it real? Strategy changes, the system doesn't. “We made the right decision, but the system won’t let it stick.”
- RENEWAL — Can we change it again? Yesterday's success becomes today's constraint. “What made us successful no longer fits, but we can’t let go of it.”
Two of these are the inputs to everything else. Energy determines how much of people’s potential actually enters the work. Truth determines how much of reality enters the organisation and is understood, including whether people are willing to say what they see. Agency, Choice, Embedding and Renewal then decide what the organisation can do with them. The six interact and reinforce one another, so this is a way to hold the whole picture rather than a maturity ladder to climb in order.
Example: the conditions give a precise diagnostic map for a specific strategic issue:
A diagnostic map for a specific strategic issue
- People had more to contribute but stopped contributing it → Energy
- The organisation never noticed the signal → Truth
- Someone noticed it but did not surface it → Truth (candour)
- It was surfaced but aggregated, dismissed or misunderstood → Truth
- It was understood, but the relevant actor could not act → Agency
- Authority existed, but incentives, targets or governance produced the wrong enterprise choice → Choice
- The right choice was made, but resources did not actually move → Choice
- Resources moved, but roles, metrics, systems and routines did not change → Embedding
- The new system worked, but later became impossible to question or replace → Renewal
The distinction matters because two organisations can show exactly the same symptom and have completely different underlying problems. Two organisations both complain that they are slow. One is slow because people are afraid to challenge senior management. Another because every decision has migrated upward. Another because it reads its market through historical data rather than emerging needs. Another because its budgeting makes the safe choice the rational one. The visible problem is the same, the causes are not.
Hybrid human/AI organisations
The same logic follows what happens across people and AI systems, because the hybrid organisation gives the conditions new ways to break: Truth can fail because a person suppresses bad news, or because an agent has incomplete data, poor retrieval, missing provenance or an incorrect model of the situation. Agency can fail because a manager lacks decision rights, or because an agent lacks permissions, tool access or an escalation path. Choice can be distorted by human incentives, or by an optimisation objective encoded into an agent. Embedding can fail because humans ignore a new workflow, or because the agent is bolted onto a legacy process rather than changing the operating system. And finally, Renewal now includes changing not just routines and KPIs, but agents, prompts, policies, models and automations that have become obsolete.
3. Trace the root cause
Finding the condition tells you where performance breaks, not why. The aim is not to keep digging indefinitely. It is to find the most upstream mechanism that is sufficiently supported by the evidence, materially explains the behaviour and is practically alterable.
That is the difference between concluding that “people need more guts” and discovering that challenging a senior decision damages performance ratings, requires three levels of approval or has repeatedly failed to change an outcome. The first points towards exhortation or training. The second gives management something concrete to change.
Underneath sit three layers. First, the deep management logic, what we implicitly believe about people, control and risk, often culturally defined. Second, the formal organisational practices and resulting unwritten rules, the behavioural logic employees apply, deciphered as: what am I trying to achieve or obtain, who has the power to give me that, and what criteria does that person or system use? Third, the reinforcing loops, how the behaviour created by those mechanisms feeds back into the system and makes the original logic appear ever more true.
An example, take overcentralisation. It looks like a root cause, but ask why. Perhaps leaders quietly believe people further down cannot be trusted to judge well. That belief hardens into approval limits and restricted information. People get fewer chances to exercise judgment, so their judgment develops less, so they escalate more, so leaders conclude they were right, and authority moves further up. That is a reinforcing loop. But loops can turn. Give people more context and they judge better; better judgment earns trust; trust moves authority down; more authority builds more judgment. The vicious loop can become virtuous. This is why the analysis cannot stop at the first mechanism it finds. It has to reach the logic that created the mechanism and the loop that keeps it alive.
Another example, take budgeting. Start with a common belief: predictability is the mark of good management. That belief becomes embedded in fixed annual budgets, forecasts that double as commitments, and investment thresholds every proposal must clear. Managers respond rationally. They lowball forecasts, protect resources and avoid uncertain opportunities that could put their targets at risk. The organisation keeps funding yesterday’s business rather than tomorrow’s.
4.Find the leverage point
The Audit is not meant to produce another twenty-page heatmap, or the conclusion “you need to transform the culture.” A conventional diagnostic reports low scores on empowerment, collaboration and innovation, then lists twelve recommendations. The Audit should be able to say something far more specific: the primary breakpoint is Embedding; the mechanism is an old KPI and incentive system that still makes the old behaviour rational; of everything observed, this has the most causal leverage; change it first and test what happens. The aim is a small, safe-to-try intervention significant enough to generate real evidence.
5. Decide what to change, protect or transfer
This is where both sides of the Adaptability Audit meet: release what holds performance back, without accidentally removing what makes the organisation work. So the closing question is what to change, what to protect, and what to transfer. Change the mechanisms constraining the performance now required. Protect the practices and unwritten rules that are quietly valuable. And where something works exceptionally well in one part of the organisation, transfer the insight into why it works, rather than copying the practice blind. This matters most in transformations and M&A, where standardising everything can make an organisation look coherent while destroying the local practices that were creating the value.
The method, a real episode, then adaptive questioning. A conventional survey asks “how empowered are people here?” The Audit asks “think of the last time an important opportunity appeared unexpectedly. What happened between someone noticing it and the organisation responding?” The first gives a perception score. The second gives an episode you can investigate, following it from who saw it, to whether it surfaced, to who could act, to whether resources moved. Each answer decides the next question, and the broad base of knowledge sits behind the conversation, deciding what to ask and when there is enough evidence to act.
What it leaves you with
For each issue it examines, the Audit should leave a leadership team with a decision-ready brief. It should be clear about what the evidence supports, where it is uncertain, and what to do next. The aim is narrow and practical: find where the current system works against the performance the strategy now requires, and change the one or two things most likely to release it.
A useful output might read:
The organisation is seeing the opportunity early enough, and local teams have the authority to test it. The constraint sits in Choice: annual resource commitments and target protection make uncertain opportunities structurally unattractive. Spain shows the alternative, through staged opportunity funding. Test that mechanism in two more markets before changing the wider allocation process, and protect the existing freedom to run low-cost local experiments.
That is worth more than knowing the organisation scored 6.2 on empowerment and 5.7 on innovation. It separates what the evidence says from what to do next and what to leave untouched, and it stays honest about where the evidence runs out and another look is needed first.
The final output therefore distinguishes between three things.
- What the evidence says. The diagnosis, supporting episodes, counterexamples and confidence in the conclusion.
- What management should do next. A small number of interventions or experiments, specific enough to act on and bounded enough to learn from.
- What should remain untouched. Existing strengths that enable performance, plus mechanisms that may be worth amplifying or transferring elsewhere.
The result is not a blueprint for an ideal organisation. It is a focused view of where management action has the greatest chance of releasing performance.
Most change efforts run diagnose, design, roll out, hope. This one runs diagnose, intervene, experience, learn.