The 10/80/10 rule

Ten percent direction. Eighty percent execution. Ten percent judgement. That is the shape of almost any piece of knowledge work, and it is the simplest map I have for where AI fits.

The first ten percent is deciding what you are actually doing and why. What good looks like. Who it is for. What it needs to achieve. This is small in time and enormous in consequence. Get the direction wrong and the other ninety percent is wasted, no matter how well it is done. A beautifully executed answer to the wrong question is still the wrong answer.

The middle eighty is the doing. The drafting, the building, the formatting, the first pass, the grind. This is where the hours go and where most people feel their work actually happens. It is also the part a machine is now genuinely good at. Fast, cheap, tireless, and good enough to be a strong starting point on most jobs.

The final ten percent is judgement. Looking at what came back and deciding whether it is any good. Catching what is wrong, what is off, what is technically fine but tone-deaf for this customer. Then putting your name to it and owning the result. Like the first ten, it is small in time and large in consequence.

Here is the mistake almost everyone makes. They think AI is taking the whole job. It is not. It is taking the eighty. The two tens stay human, because they are the parts that need to know your business, your customer, and your standards, and the tool knows none of those. When people hand the tool one of the tens, things go wrong fast. Let it set the direction and you get generic work aimed at nobody. Skip the judgement and you ship the mistakes at speed.

I watched a small consultancy learn this. They started using AI to write client reports and, for a giddy month, thought they had automated the whole thing. Quality dropped. Clients noticed. The reports were fluent and hollow. The fix was not to drop the tool. It was to put the tens back. A human decided what each report needed to say and why, that was the first ten. The tool produced the bulk, that was the eighty. A human read it properly, sharpened it, and stood behind it, that was the last ten. The reports got better than before, and they took less time, because the humans spent their energy only where it counted.

The sharper way to use this rule is as a question you ask on every task. Where is the ten percent of direction here, and have I done it. Where is the ten percent of judgement, and who owns it. If both tens are covered by a person who cares, give the eighty to the tool with confidence. If either ten is missing, no tool will save the work.

This also tells you where to spend your own time. Your judgement is your most valuable asset, and the eighty percent is the cheapest part of the job to get help with. So an owner doing the middle by hand is spending the expensive thing on the cheap work. The whole point of the rule is to move your attention out of the eighty and into the tens, where it actually changes the outcome.

The practical takeaway. Take your next real task and split it into the three parts before you start. Decide the direction yourself. Hand the bulk to the tool. Then check and own the result yourself. Protect the two tens like they matter, because they are the whole reason the work is yours and not just anyone’s.