How much of each healthcare job can AI do?
Search your title or pick it. The headline reads your job the way AI meets it, task by task. Open a job to see the frontiers it draws on, how good AI is at each today, its best models and the real tests behind them, and your job's actual tasks under each one, every task marked by how much AI can do it.
This shows what AI can do, not that your job goes away. The hands-on and in-person parts mostly stay. Jobs built around talking and analyzing, some physicians included, can read higher than feels right; open the tasks to see what is driving the number.
All 923 jobs are listed below, healthcare first. Type above to jump to yours, then open it for the full breakdown.
No job matches that. Try a shorter word, like "care" or "sales".
No, not the job. About 48 percent of a registered nurse's work is the kind AI can take on today, and it is the documentation and information side. The assessment, hands-on care, and reading a patient in the room stay with people.
The most exposed are records and text heavy roles: billing clerks around 66 percent, receptionists around 61 percent, coders and secretaries around 57 percent. The least exposed are hands-on: home health aides around 20 percent, dentists around 33 percent.
Coding and billing are among the most exposed healthcare roles, roughly 57 to 66 percent, because most of the work is structured text and data. That is not replacement; it shifts the role toward exception handling, audit, and judgment.
About 61 percent. AI can track inventory and explain medications, but the safety check, catching a dangerous interaction before it reaches the patient, stays with the pharmacist.
Across nursing, therapy, aide, and physician roles, this analysis marks the same core as human: hands-on clinical care, physical exams and procedures, and being in the room with patients.
Want a deeper write-up for a specific role, or exactly how these numbers were built?
Read the method › · See how we think about AI for your practice ›