Deciding What to Automate Introduction

How to tell which work is worth handing to an agent

Four things you can do with a task, three questions in the order that decides, and why the shape that looks worst is usually the one to augment.

Part of the leadership bundle · coming soon
You can read this introduction now. The 12 steps open when purchases do. Meanwhile, The Judgment Gap is free to read in full.
Part of the leadership bundle: $49.99
Every leadership volume, including new ones when they're released. One-time purchase, free updates. The Judgment Gap is free to read in full.
Decision Q&A

Four questions this volume answers

1Why does it matter to me?

You're buying the doing minus the checking, and checking didn't get cheaper because it needs somebody who knows the domain. So a proposal can have the arithmetic right and still lose money. And the choice isn't automate or not: it's one of four, automate, augment, remove or leave alone.

Shown in step 1 · 2 · 3
2What should I ask?

In this order: how cheaply can somebody establish that the result is right? How often does it happen? How much does a wrong answer touch, and how long before anybody notices? Then add four rows to the proposal: who checks it and how long it takes, who notices when the surrounding system changes, how we'd find out it quietly started being wrong, and whose name is on it in a year.

Shown in step 4 · 5 · 11
3How will I know it's going wrong?

A person still reading every output weeks in. Nobody able to say how you'd find out it had started being wrong. A trial with no stated result that ends it, no date and no named person, which drifts instead of ending. And an automated step whose output nobody reads, which is now permanent.

Shown in step 6 · 10 · 11
4What do I decide?

One disposition per task. A machine can check it: automate. It's expensive to check because a second person would rebuild the context: augment, so the person keeps doing it with the agent helping. Nobody needs its output: remove it, and test that by stopping it for a month. Otherwise leave it alone. And don't discard low-volume work that's cheap to check just because volume was your first filter.

Shown in step 6 · 8 · 9 · 10 · 12

There are four things you can do with a task, not two.

What you are buying is not the doing. It is the doing minus what it takes to find out whether the result is right — and that second cost is the one that did not fall, because checking needs somebody who knows the domain well enough to recognize a wrong answer.

Which means the decision is not automate-or-not. It is one of four, and how cheaply you can check the result decides which. This volume is twelve short steps on getting that assignment right. It assumes you have read nothing else in this series.

one column collapses. The total barely moves doing it checking it what it costs you collapses unchanged and this is what you buy the proposal quoted the first column and called it the saving. every number in it is right. the arithmetic is not the error.
The spine of the volume: automation collapses one of the two costs and leaves the other where it was, so the total falls far less than the proposal claims.
What moves: the cost of doing the work collapses while the cost of checking it does not move at all, so the total falls only a little.

The four

the dispositionwhat it means
automatethe agent does it, and you check the result cheaply — or a machine does
augmentthe person keeps doing it and the agent helps; the checking rides along with the doing
removethe step should not exist. Automating it would make the workflow permanent
leave alonenone of the above is worth the change. The default, and it needs saying

The most common wrong answer in this subject is no given to a task that should have been augmented. High volume and expensive checking looks like the worst possible candidate, and it usually means the person should keep doing the work with the agent helping — because then nobody has to rebuild the context, and the checking cost does not arise at all rather than merely shrinking.

And the highest-leverage page is not about agents. Some steps on your list should not exist. Automating one of those makes it permanent, because now a schedule and a downstream job depend on it — so it becomes harder to delete than it was when somebody was doing it grudgingly by hand.

What you can do afterward

Read a proposal for the cost it left out. Ask the one question that gets you that cost. Assign one of four words to every candidate on your list and defend it. And add four lines to any proposal you are handed, answerable by people already in the room.

The four dispositions, the three questions and the four missing costs are on one page you can print.