TL;DR
Two businesses buy the same tool, run the same implementation, and get different outcomes. The variable is organisational capability, which is the practical ability of a business to act. Capability loses arguments against technology proposals because a licence cost is a number and capability usually feels like an opinion. It can be counted. Six measures make it observable, none needs a platform or a quarter of analysis, and each is compared against itself over time so no benchmark is required. Naming this as a discipline matters, because things without a name rarely get an owner or a budget.
The most useful thing I have seen in a technology review was a slide nobody had planned to present.
Two divisions of the same group had rolled out the same tool, on the same contract, with the same implementation partner and the same eight weeks of training. One had adoption above eighty per cent and a measurable reduction in cycle time. The other had adoption in the low thirties and no measurable change at all.
Everything the business case had modelled was identical. The outcomes were not.
This happens often enough to be the normal result, and it rarely gets investigated properly, because the investigation would have to look somewhere uncomfortable. The variable is not the technology. It is the organisation the technology landed in.
What capability actually means
Capability is the practical ability of a business to act. The test is whether an intended action reliably happens.
Treating it as a skills question is the most common misunderstanding, and the one that wastes the most training budget. Knowing how to do the work is part of it. The rest is the conditions around the person: whether they have the time to do it properly, the authority to act on what they find, and whether the information, tools and management around them make the right action the easy one.
A weakness in any of those produces the same visible symptom. Someone does not follow the new process. The business concludes they need training, buys more training, and the symptom returns six weeks later because the real constraint was that the person could not approve their own decision and the approver was in another time zone.
Capability is also built by subtraction. A team becomes more capable when a process gets simpler. A manager becomes more capable when a dashboard becomes trustworthy. Neither requires a budget, which is one reason neither gets proposed.
Why the argument gets lost
Capability tends to lose arguments against technology proposals for a structural reason, not an intellectual one.
A licence cost is a number. A day rate is a number. A projected saving is a number, even when the assumptions behind it are soft. Capability arrives in the same meeting described as readiness, culture or maturity, and none of those can be argued with or acted on. Words that cannot be disagreed with cannot be prioritised either.
So the technology proposal wins, not because it is better reasoned, but because it is the only one carrying figures.
The way out of this is not a better argument. It is a number.
Six measures
None of these produces an absolute score. Each produces a figure you can compare with the same figure eight weeks later, which is enough to show whether capability is improving or slipping. Consistency matters more than precision, and all six can be gathered from inside the business without a platform.
Escalation rate on a named decision
Pick one decision that gets escalated. Count how many escalations the rules actually required. The gap between the total and the required number is the part being escalated because nobody is confident enough to decide, which is an authority and confidence measure wearing a process disguise.
Return rate at a single handover
Pick one point where work crosses a boundary. Count what proportion comes back for missing or unclear information. Handovers are where most organisational friction lives and where almost nobody looks, because the stages either side have owners and the gap between them does not.
Correction rate on AI output
Of the AI-generated content that gets used, what proportion needs substantive change before it goes anywhere. This is the measure most likely to be resisted, and the most directly connected to whether the investment is working.
Time from work-ready to decision made
Measure the gap between work being ready for a decision and the decision happening. Compare it to the time the work itself takes. In most workflows, waiting is the larger number, and automating the doing has a low ceiling while that remains true.
Concentration
What share of a workflow’s output comes from the two most experienced people. A high figure means the process is carried by judgement nobody has written down. It is also a risk measure, and framing it that way tends to get it funded faster.
Definition variance
Ask six people across two teams to write down what a key term means. Count the distinct answers. This one takes ten minutes, needs no preparation, and is the most reliably uncomfortable of the six.
The two places capability usually breaks
Across enough of these, two constraints come up far more often than the rest.
The first is authority. Someone can see what needs to happen and cannot do it without asking. The asking is usually quick, the waiting is not, and over a few hundred instances it becomes the dominant cost in the workflow. It is rarely visible in any report, because the time sits between two steps that both look efficient.
The second is definition. Two teams use the same word to mean different things, and every process built on that word inherits the ambiguity. Qualified, complete, resolved, active. The disagreement is invisible until something forces the two definitions into the same report, at which point the conversation becomes about the numbers instead of the decision.
Neither is expensive to fix. Authority is usually a sentence in a role description and a manager saying out loud that they meant it. Definition is a meeting and a written answer. Both are cheap, both are slow to get prioritised, and both cap the return on a six-figure technology investment.
That mismatch, between the cost of the constraint and the cost of the thing it constrains, is the practical argument for measuring capability at all.
What the numbers do
The point of measuring is not the measurement. It is that capability stops being an opinion and starts behaving like everything else the business manages.
A figure taken twice becomes a direction. A direction can be reported. Something that can be reported can have an owner, and an owner can be asked about it in the same meeting where the licence cost gets discussed. That is the whole mechanism, and it is more mundane than it sounds.
It also changes what happens when a technology investment underdelivers. Without a capability measure, the conversation is about whether the tool was the right one. With a baseline taken before deployment, the conversation is about which specific constraint capped the benefit, and that conversation leads somewhere.
Why this needs a name
Things without a name rarely get an owner or a budget.
Transformation has a name, so it gets a programme, a director, a budget line and a place on the board agenda. The organisational conditions that decide whether any of it lands get a paragraph in the risk section.
Edge151 uses Business Evolution for the named version: the deliberate, continuous building of organisational capability at the rate a business’s ambition requires. Deliberate means designed instead of assumed. Continuous means it has a rate instead of a completion date. The definition is unpacked properly on its own page, and the point of naming it is narrow and practical. A named discipline can have an owner, a method, a set of measures and a rate. An unnamed one gets a paragraph.
Where to start
Pick the workflow where a symptom keeps returning and where improvement would matter commercially. Run two of the six measures against it. Write the numbers somewhere you will look again in eight weeks.
That is not a programme and it does not need approval. It is two numbers and a date in the calendar.
The comparison, taken twice, is a capability measure. It is also the smallest possible piece of evidence that the gap between what a business has spent and what it is getting back can be closed on purpose.
Take one workflow where a symptom keeps returning. Run the definition variance measure on it, which takes ten minutes, and the return rate at one handover, which takes a week of counting. Put both numbers in your calendar with a date eight weeks out. What you do next depends on which direction they move.
Series navigation
This is the first of five articles setting out the case for Business Evolution. Each stands on its own, and they build in order.
Previously: Your AI business case assumes a capability you may not have. The four organisational assumptions buried in the benefit line, and six tests to run before approving one. Read it here
Next: Why transformation programmes keep failing the businesses that buy them. The structural reason programmes decay after closure, what the restart cycle costs, and what a sustainable rate requires instead. Read it at coming soon.
The full definition sits on the pillar page: What Business Evolution is.
The five
- Your AI business case assumes a capability you may not have.
- Capability is the connective tissue between AI spend and business value. You are here.
- Why transformation programmes keep failing the businesses that buy them.
- Three altitudes of change: organisation, workflow, task.
- What a management discipline needs before it earns the name.
Organisational capability is the practical ability of a business to act, tested by whether an intended action reliably happens. It is broader than skills. It includes whether people have the time and authority to act, and whether the information, tools and management around them make the right action the easy one.
Because the tool is the constant and the organisation is the variable. Workflow clarity, decision ownership, data that a decision actually depends on, and whether people have the authority to act on what the tool produces all differ between organisations, and all of them cap the benefit independently of the technology.
Six measures cover most of it: escalation rate on a named decision, return rate at a single handover, correction rate on AI output, time from work being ready for a decision to the decision being made, the share of a workflow’s output produced by the two most experienced people, and how many distinct definitions exist for a key term. Each is compared against itself over time.
No. External benchmarks invite arguments about comparability and delay the work. Each of the six measures is compared against the same measure taken in the same business eight weeks earlier. The change between two readings is the useful information.
Culture describes how people behave. Capability describes what the organisation makes possible. Most behaviour that gets labelled cultural turns out to be structural, because people generally do whatever the system makes easy. That is why capability work often produces a change in behaviour that a values programme could not.
Alastair Jupp writes on organisational capability and the practical adoption of AI. The argument here is developed at length in Workflows, Decisions, Discipline: The Operating System Behind Every High-Performing Business, published by Edge151 later this year. Details and publication updates.
Discover more from Edge151
Subscribe to get the latest posts sent to your email.