Scheduling is the most exposed thing anybody does at work

The index scores 2,087 standardised work activities, grouped into 41 families. Ranked by mean exposed share, the top of that list is not close.

Scheduling scores twenty points clear of the next family. It is the single most exposed category of work in the index.

Why it scores the way it does

Scheduling has an unusual profile. Capability is at the top of the scale, because the work product is a plan expressed as data, and systems produce those well. Every friction is near the bottom. There is nothing physical about it, nobody needs to be present, no licence attaches, and an error announces itself immediately when somebody turns up at the wrong time. Cheap verification, no liability, no body required.

That combination is rare. Most highly capable work carries at least one serious friction. Scheduling carries almost none.

The part that matters is where it lives

Scheduling is not a job for most of the people who do it. It sits inside other jobs, and the fourteen activities in this family attach to occupations across nearly every major group: managers, administrators, nurses, supervisors, teachers, dispatchers, project leads.

This is the argument for task-level measurement stated as plainly as the data ever states it. An occupation-level index would show a practice manager as moderately exposed and leave it there. The task-level view shows something more actionable: a specific, bounded, universally present category of work that is almost entirely automatable, sitting inside hundreds of roles that are otherwise well protected.

For most people the honest answer to "what will AI take from my job" is not the job. It is this, plus the documenting and recording that sits one row below it.

What the low end looks like

For contrast, the bottom of the same ranking:

Repairing mechanical equipment scores zero across all 32 of its activities. Not low. Zero. No activity in that family is rated as something a current system can produce, because every one of them requires acting on a physical object.

The gap between the top and bottom of this table is the whole story of this wave of automation, and it has almost nothing to do with how difficult the work is or how long it takes to learn.