Automatable

How automatable is your job?

Pick your occupation to get your automation score, then the exact tasks worth handing off and how to do each. Built on U.S. Department of Labor task data and a Stanford study of 1,500 workers.

Try “Web Developers”, “Accountants”, or “Mechanical Engineers”, or browse all 104 occupations.

  1. 1Get your automation scoreSee how automatable your job is, across 104 occupations.
  2. 2See what to hand offThe tasks in your role worth automating, ranked.
  3. 3Start with the right toolsThe no-code and AI tools, and how to do each.
104
occupations
844
work tasks rated
56%
of tasks workers want automated
33%
ready to automate now

Key statistics

Free to cite, with a link back to Automatable.

  • Across 104 U.S. occupations, workers want AI to automate about 56% of their work tasks (Stanford WORKBank, 1,500 workers).
  • 33% of the 844 studied tasks (281 tasks) are "Automation Green Light": workers want them automated and AI can already do them (Stanford WORKBank).
  • Medical Transcriptionists is the most automatable occupation studied: AI can already do 100% of its 8 tasks (Stanford WORKBank, O*NET SOC 31-9094.00).
  • Art Directors is among the least automatable: AI can do only 0% of its tasks today (Stanford WORKBank, O*NET SOC 27-1011.00).
  • The median studied task scores 3/5 for worker desire to automate and 3.5/5 for AI capability (Stanford WORKBank, 1-5 scale).
  • Automatable maps 844 real O*NET tasks across 104 occupations; source: Stanford WORKBank (Shao et al., 2025, arXiv:2506.06576).

Two questions decide every task

For each task we cross what workers want automated with what AI is able to do. That puts every task in one of four zones.

The most and least automatable occupations

Ranked by the share of each role’s tasks that AI experts rate as do-able today, the canonical Automatable score, across all 104 occupations in the Stanford WORKBank study.

Built on real evidence

Every figure comes from the Stanford WORKBank study (Shao et al., 2025), which surveyed 1,500 U.S. workers and 52 AI experts about 844 real occupational tasks from the U.S. Department of Labor’s O*NET database. We don’t guess, we report what workers said and what experts rated.

How we built this →