Intelligence work, judgement work
Before choosing a tool, you need to know what kind of work is on the table. That separation decides where AI lands first and where it should not land yet.
- 01 · You are here
- 02 · Where you already outsource
- 03 · Why your pilot never shipped
- 04 · Who answers when AI gets it wrong
- 05 · The agenda your budget holder approves
Classify any activity in your operation between repeatable intelligence and judgement, and use that position to decide what is worth automating now, what is worth assisting and what stays human.
The wrong question costs the whole budget
The conversation almost always starts in the same place: which tool to buy. The choice becomes a feature comparison, a price per user and a rollout timeline. Months later the licence is paid, a few people use it, and nobody can say which work is now done differently.
The reason is simple. A tool is an answer. Nobody had defined the question: which work, done one way today, no longer needs to be done that way.
A complex rule is not the same as judgement
Turning a specification into code, checking a standard contract against a clause list, classifying an invoice, reconciling a bank statement. These take real expertise and cost money when done badly. But the rules governing them, however complex, are still recognisable as rules. They can be written down, reviewed and verified.
Now compare: deciding what to build next quarter, accepting technical debt in exchange for a deadline, granting an exception to a supplier who missed a date but matters strategically. There is no list here. There is judgement built through practice, context and responsibility for consequences.
The larger the share of repeatable intelligence in an activity, the more likely the current way of doing it gets reorganised by AI systems. The share is what matters, not the profession.
A contracts analyst spends part of her time checking clauses against a standard and part negotiating exceptions with internal teams. The first share is highly repeatable. The second is pure judgement. The activity does not leave her desk: what changes is how much time she spends on each half.
Reading this distinction as a list of who leaves leads to the wrong decision. It exists to find where a person is being used as if they were a rule, and where their attention can move to exceptions, decisions, creation, negotiation and responsibility.
Selling a tool means competing with the next model
When what you deliver is a tool, every new model threatens the proposition. When what you deliver is completed work, every model improvement makes the delivery faster and more controllable.
The practical consequence is economic. The budget attached to a profession's work tends to be larger than the budget for software per user. That is the logic behind charging for operational capacity used, rather than for an active licence.
If your next decision is which subscription to buy, you are buying a tool. If it is which work gets delivered differently, and who answers for it, you are redesigning an operation. The two decisions cost about the same and produce very different results.
“Our operation is far too complex for this”
It probably is. But complexity is rarely evenly distributed. Inside any complex operation there are activities whose rule is long, dull and stable. Those tend to absorb hours of expensive people and resist every attempt at improvement, because they were never treated as a design problem.
The exercise below is not there to classify the company. It classifies five concrete activities and shows where they land.
Which part of your judgement is still implicit
If part of your team's work is repeatable, part of yours is too. Approving what has already been approved ten times against the same criteria is not judgement: it is a rule nobody wrote down.
The uncomfortable question is how much of your own criteria you could make explicit so someone else could decide in your place. What remains after that is the judgement that stays yours, and the responsibility that stays human.
Classify five activities in your operation
Pick five recurring activities that consume senior people's time. Position each between repeatable intelligence and judgement. Do not chase precision: chase the relative order between them.
Ruler: written rule or judgement
Use the controls or the arrow keys. Zero means work almost entirely governed by an explicit rule. One hundred means work that depends on criteria built through experience.
Your entries are stored on this device only and are not sent to TheNeil. If you clear browser data or switch devices, that record may be lost. Export and print are how you keep the result.
Take the two left-most activities into your next meeting
Take the two activities closest to zero and answer three questions with whoever executes them: how many hours a week they consume, who reviews the output and what happens today when they come out wrong. Those answers are the input for course 2.
Where the argument comes from
- Julien Bek, “Services: The New Software”, Sequoia Capital, 2026. Supports the distinction between selling a tool and selling completed work, and the comparison between services budget and software budget.
- Autor, Levy and Murnane, 2003. Supports the idea that activities governed by explicit rules are the first to be reorganised by systems.
When the classification becomes an investment decision
Prumo Discovery is TheNeil's paid diagnosis: fixed scope and timeline, run with your teams, and it can recommend not proceeding. It is the next step when there is a candidate operation and the decision to invest is still open.