01
The measurement ships with the product
Capacity released and redeployed appears as an operating metric in the product itself, not as a report we generate for a case study. If it is only visible when we are selling, it is marketing.
Human Opportunity Initiative · Capacity programme
The industry measures the work AI removes. Nobody measures what replaced it.
This is a thesis we are building on and a framework we are developing, not a result we have proven. We have no deployments yet. Everything here is labelled accordingly.
The argument
Automation does not destroy capacity. It relocates it.
When an AI system takes over a repetitive process, something is produced as well as removed. Hours come back. Attention comes back. A team that spent nine hours a week reconciling records has nine hours it did not have before.
That capacity is a real asset, and it is the only part of the transaction that almost nobody puts a number against. The industry reports the hours removed, because that figure is easy to calculate and flattering to the vendor. What happened next is harder to measure and far more consequential.
Automation creates capacity. Most companies spend it once, as a headcount line. It can be invested instead.
Both are legitimate business decisions and we are not going to pretend otherwise. But they have very different shapes over time. Cut capacity is a single-year saving that does not recur. Redeployed capacity compounds, and it keeps the institutional knowledge that took a decade to build.
The reason this rarely gets decided properly is not that executives are careless. It is that one side of the trade has a number and the other does not. You can put "220 hours removed" in a board pack. You cannot currently put anything credible next to it.
So we are trying to build the other number. Not because it is virtuous, but because a decision made with half the information is usually the wrong decision, and this one is being made at scale right now across every industry we work in.
The model
Five steps, and only one of them is optional.
Steps one and two happen automatically whenever automation works. Step three is a choice, and it is the one the whole thesis turns on.
-
01
Automate
AI absorbs the repetitive work — the chasing, reconciling, re-keying and reporting.
Happens automatically
-
02
Capacity is released
Hours return to people. Usually fewer than the headline claim, because the new system consumes some itself.
Happens automatically
-
03
Capacity is redeployed
Those hours are deliberately directed into judgement, creativity, relationships, strategy and problem solving.
A decision. The hinge.
-
04
Opportunity is created
New skills, changed roles, responsibilities that did not previously fit into anyone's week.
Follows from step three
-
05
The business compounds
Revenue, quality, innovation and customer experience improve, because capable people have room to do work only people can do.
Follows from step four
Steps one and two are engineering. Step three is a management decision, made once, usually quickly, and usually without a number in front of it. Everything the initiative is trying to do is aimed at that single moment.
The Human Opportunity Index
A measurement we are building, in the open.
Draft framework · v0.1 · Not validated
If the argument is that redeployed capacity matters more than removed work, then it has to be measurable, or it is just a nicer way of saying the same marketing. This is our first attempt at that instrument.
Redeployment Rate Capacity redeployed Capacity released
The single number the index resolves to. A programme that automates a great deal and redeploys almost none of it scores badly here, which is the entire point.
-
01
Work automated
Hours of repetitive work removed
The only number the industry currently reports. On its own it says nothing about outcome.
-
02
Capacity released
Net hours returned to people
Automated hours minus the hours the new system itself consumes. Usually smaller than the headline figure.
-
03
Capacity redeployed
Released hours moved to higher-value work
The hinge of the whole framework. Released capacity that goes nowhere is not an outcome.
-
04
Work elevation
Share of redeployed hours in judgement work
Distinguishes moving people to better work from moving them to different repetitive work.
-
05
Capability growth
Skills acquired, roles changed
Whether the organisation actually invested in people being able to do the higher-value work.
-
06
Opportunity created
New roles and internal mobility
Positions and responsibilities that did not exist before the capacity appeared.
-
07
Value return
Business value attributable to redeployed capacity
Closes the loop. Without this the framework is sentiment rather than measurement.
What we are committing to
Four commitments that cost us something.
A thesis with no cost attached is a marketing position. These are the constraints that make this one different.
02
We publish the methodology, including revisions
The index definitions will change as they meet real deployments. We intend to publish those changes rather than quietly restate the framework, so anyone can see where we were wrong.
03
We will report the uncomfortable number
If a deployment releases significant capacity and the organisation redeploys none of it, that is a low score and we will say so. A measurement that only ever flatters the vendor is not a measurement.
04
We will not claim a result we have not measured
Until an actual engagement produces an actual number, this page says thesis. That constraint costs us persuasive power now and it is the only thing that will make the number mean anything later.
The objections
The five things people say about this, answered.
Most of them are reasonable, and one of them is the reason we might be wrong.
Isn't this just a nicer way of saying you will cut jobs?
No, and we are not going to pretend the question is unfair. Some roles change and some tasks disappear — that is what automation does. Our argument is about what happens to the capacity that creates. A company that treats it purely as a headcount reduction takes a one-year saving and loses knowledge that was going to compound. A company that redeploys it keeps both. We think the second is the better business decision, and we want the measurement to exist so it can actually be made rather than assumed.
Why would a company care about this if the cost saving is already banked?
Because the saving is banked once. Capacity that goes back into customer relationships, quality, or the work nobody has had time to do keeps producing. The uncomfortable version of the argument is that most organisations cannot currently tell you which of the two they did, because nobody measured. That is the gap we are trying to close.
Is this corporate social responsibility?
No. There is no fund, no pledge and no cause. It is an operating measurement and a product decision — the capacity number sits on the executive dashboard next to revenue and cash, not on a values page. If it were CSR we would have put it in the footer.
Are you anti-AI?
We build AI systems. The position is not that automation is bad; it is that measuring only the work removed is an incomplete way to evaluate an AI system, and the industry has settled on it because it is the easy half.
Has any of this been proven?
No. We have no deployments, and the index has never been used to score an organisation. Everything on this page is a thesis and a framework, and we have tried to label it that way throughout. If we had evidence we would show it.
Next: The Human Opportunity Pipeline — the designed pathway from graduate to professional. Both sit under the Human Opportunity Initiative.