The problem I had
I find myself fighting in sessions to get AI to understand the end goal. What I want is an obsession with executing activities that produce value. Nick, 2 September 2026
I could describe my business in a sentence and I could describe what I did yesterday. What I could not do was put the two on the same page. Every tool I tried wanted a process diagram, and a process diagram answers a question nobody in a small company is asking.
So I drew a grid instead. Eight functions, because a company of any size is doing eight kinds of thing whether or not it has departments for them. Eight rows of eight, and a rule that every square gets filled in with who does it and whether it runs. Sixty-four is not a magic number. It is small enough to finish in an afternoon and big enough that you cannot bluff it.
What the picture shows
Ferncastle does fifty-two of the sixty-four. Twelve squares are empty, which is its own finding: three of the empty ones are in Site surveys, the part of the business its customers actually pay for.
Eighteen jobs genuinely need a person. The technician deciding whether a plant is coming back. The surveyor reading light at a window. The chief executive signing anything over two hundred plants. Those are not automation debt, and a chart that treats them as debt is lying to you.
That leaves thirty-four jobs with no good reason to have a person on them. Sixteen of them run. Eleven of those sixteen have been seen working; four can actually prove it; and one has been off for eighteen days while everyone assumed it was fine.
The remaining eighteen are the real finding, and they are the ones that look like nothing on a list. Eight are done by hand every week by somebody competent, which is exactly why nobody escalates them. Ten were never built. Read as a list they are unremarkable. Read as eighteen hollow squares next to sixteen filled ones, they are the company's actual position.
What changed
One rule: every square names a performer and a state, and "a person does it" is a state, not a blank. The moment that rule existed, the argument about what to automate next stopped being a matter of opinion. You point at the outlined squares.
The second thing it changed was subtler and mattered more. Press the goal switch above and the picture fills in. That is not a target for next quarter; it is the shape of the company when the work that never needed a person stops having one. The grey squares do not move. Seeing what stays human is the part that makes the rest of it safe to do.
The reason to draw it rather than write it is that a list of fifty-two jobs reads as a company doing well. Thirty-four squares with eighteen of them hollow does not.
The four things I weighed, and why this one
- To start
- An afternoon and one sheet. The fixed size is what forces the finishing.
- What it fixes
- The empty squares are as informative as the full ones, and neither can be talked away.
- What it costs
- Some jobs do not fit their function cleanly, and you have to put them somewhere anyway.
- Why this one
- It is the only version where "we have not built that" and "we do not do that at all" look different on the page.
- To start
- Weeks, and a consultant.
- Why not
- It answers how work moves between steps. The question here is which steps have a person on them and why, and a flow diagram buries that under arrows.
- To start
- A morning with a spreadsheet.
- Why not
- A ranked list has no bottom, so nothing is ever finished and the jobs that should stay human sit permanently near the end of it looking like failures.
- To start
- Nothing, which is the warning.
- Why not
- Without a fixed size there is no such thing as an empty square, and the empty squares are half the value. A grid that grows to fit is a list wearing a grid's clothes.