Once roles are broken into tasks, each task gets one of three labels. The label drives both the capacity math and the people plan.
Automate
AI performs the task with little human input: routine questions, data entry, scheduling, standard reports. Most of the time is freed, but someone still monitors quality and handles exceptions.
Augment
A person does the task with AI's help: drafting, research, analysis, preparing decisions. Time is saved and quality can rise, but the person stays in charge. Skills change: people need to prompt, check and decide.
Keep human
Tasks that need judgment, empathy, trust, negotiation or accountability: coaching, sensitive cases, hiring decisions, conflict, leadership. Little or no time is freed — and these tasks often become a larger share of the role.
Why the difference matters for people
Researchers at Stanford's Digital Economy Lab found that employment of young workers (ages 22–25) in the most AI-exposed occupations is now about 19% below where it would have been had it kept pace with less-exposed occupations — and that the declines are concentrated in occupations where AI usage primarily substitutes for human tasks, while where it complements workers, employment is flat or rising.
So automation-heavy roles need redeployment plans; augmentation-heavy roles need reskilling; human-heavy roles need protection of the time that matters.
How to set the labels
- Start with a simple rule (for example, routine and rules-based → automate; drafting and analysis → augment; care and judgment → keep human), then adjust task by task with the people who do the work.
- Treat time-freed shares as assumptions until a pilot measures them.
- Re-label every six to twelve months; AI capability changes quickly.