AI in healthcare: which tasks to automate first
Updated
Across the 19 million US jobs in healthcare, the work scores 38 of 100 for AI exposure, weighted by how many people do each job. AI can take on about 657 hours a year per full-time person, 33% of a working year. The task that frees the most hours across the industry is to maintain accurate, detailed reports and records, about 72 hours a year for each person who does it.
- AI exposure, weighted by employment
- 38 out of 100 Moderate exposure
- Hours AI can take on
- 657 hours a year per full-time person
- Task time AI can do today
- 14% AI speeds up another 47%
Tasks to automate first
The 15 tasks where AI can take on the most hours across healthcare: how exposed each task is, how much of the job it takes and how many people do that job. Hours are per person in the role that does the task. How we score each task.
| Task and who does it | AI exposure |
|---|---|
| Maintain accurate, detailed reports and records. Documentation is text heavy; ambient scribes and drafting tools fit well. Done by nurses. | 70% AI can do it 72 hours a year for each person in the role |
| Record patients' medical information and vital signs. Devices capture vitals; charting and data entry increasingly automated. Done by nurses. | 45% AI helps 47 hours a year for each person in the role |
| Consult and coordinate with healthcare team members to assess, plan, implement, or evaluate patient care plans. Care plan drafting and team messaging benefit; coordination stays human. Done by nurses. | 40% AI helps 40 hours a year for each person in the role |
| Instruct individuals, families, or other groups on topics such as health education, disease prevention, or childbirth and develop health improvement programs. Health education content can be drafted; live teaching stays human. Done by nurses. | 40% AI helps 38 hours a year for each person in the role |
| Schedule and confirm patient diagnostic appointments, surgeries, or medical consultations. Scheduling and confirmation run well on automated tools. Done by medical secretaries. | 75% AI can do it 119 hours a year for each person in the role |
| Monitor, record, and report symptoms or changes in patients' conditions. Bedside observation stays human; alerts and charting get automated. Done by nurses. | 35% AI helps 36 hours a year for each person in the role |
| Answer telephones and direct calls to appropriate staff. AI voice agents can answer and route calls today. Done by medical secretaries. | 65% AI can do it 106 hours a year for each person in the role |
| Work with individuals, groups, or families to plan or implement programs designed to improve the overall health of communities. Program planning and materials can be drafted; community work is in person. Done by nurses. | 35% AI helps 31 hours a year for each person in the role |
| Complete insurance or other claim forms. AI drafts claim forms from chart data for review. Done by medical secretaries. | 65% AI can do it 103 hours a year for each person in the role |
| Receive and route messages or documents, such as laboratory results, to appropriate staff. Rules and AI triage can route results and messages. Done by medical secretaries. | 65% AI can do it 102 hours a year for each person in the role |
| Monitor all aspects of patient care, including diet and physical activity. Monitoring diet and activity is mostly observation; logging can be assisted. Done by nurses. | 30% AI helps 29 hours a year for each person in the role |
| Modify patient treatment plans as indicated by patients' responses and conditions. Treatment changes need clinical judgement; AI offers decision support. Done by nurses. | 30% AI helps 29 hours a year for each person in the role |
| Operate office equipment, such as voice mail messaging systems, and use word processing, spreadsheet, or other software applications to prepare reports, invoices, financial statements, letters, case histories, or medical records. LLMs and office software speed report and document preparation. Done by medical secretaries. | 60% AI helps 93 hours a year for each person in the role |
| Perform bookkeeping duties, such as credits or collections, preparing and sending financial statements or bills, and keeping financial records. Billing software and AI handle statements, staff handle exceptions. Done by medical secretaries. | 60% AI helps 89 hours a year for each person in the role |
| Maintain medical records, technical library, or correspondence files. AI can index and classify records, staff still oversee filing. Done by medical secretaries. | 55% AI helps 86 hours a year for each person in the role |
What stays with people
39% of the task time across these roles needs a person, and the output of the rest still needs someone to check it. The tasks below take the most of that time.
- 14% AI can do it
- 47% AI helps
- 39% Needs a person
- Administer medications to patients and monitor patients for reactions or side effects. Nurses, 8% exposed
- Provide health care, first aid, immunizations, or assistance in convalescence or rehabilitation in locations such as schools, hospitals, or industry. Nurses, 6% exposed
- Conduct specified laboratory tests. Nurses, 15% exposed
- Observe nurses and visit patients to ensure proper nursing care. Nurses, 15% exposed
- Prepare patients for and assist with examinations or treatments. Nurses, 5% exposed
Adoption path
Start with the tasks where AI helps a person, then automate the routine ones, then give whole workflows to agents with a person approving the output.
Teams in this industry
The business functions in healthcare, by share of the workforce, with each one's exposure.
Roles in healthcare
The 20 largest roles, each linked to its task-by-task score. Shares come from the BLS industry staffing estimates, May 2025. Of the roles with at least 2% of the jobs, medical secretaries score highest at 65, nursing assistants lowest at 16.
- Nurses, 17.4% of the workforce, score 36 out of 100
- Nursing assistants, 7.9% of the workforce, score 16 out of 100
- Medical secretaries, 5.3% of the workforce, score 65 out of 100
- Medical assistants, 4.7% of the workforce, score 32 out of 100
- Licensed practical nurses, 3.3% of the workforce, score 22 out of 100
- Healthcare managers, 2.7% of the workforce, score 50 out of 100
- Receptionists, 2.3% of the workforce, score 62 out of 100
- Dental assistants, 2.3% of the workforce, score 26 out of 100
- Nurse practitioners, 1.8% of the workforce, score 49 out of 100
- Physical therapists, 1.5% of the workforce, score 38 out of 100
- Office clerks, 1.4% of the workforce, score 60 out of 100
- Administrative supervisors, 1.3% of the workforce, score 55 out of 100
- Maids, 1.3% of the workforce, score 8 out of 100
- Dental hygienists, 1.3% of the workforce, score 20 out of 100
- Radiologic technologists, 1.3% of the workforce, score 29 out of 100
- Food servers, 1.2% of the workforce, score 23 out of 100
- Billing clerks, 1% of the workforce, score 59 out of 100
- Physician assistants, 0.9% of the workforce, score 30 out of 100
- Customer service representatives, 0.9% of the workforce, score 74 out of 100
- Cafeteria cooks, 0.9% of the workforce, score 25 out of 100
Questions about AI in healthcare
Which tasks should a healthcare provider automate first?
The first task to automate is to maintain accurate, detailed reports and records (70% exposed). Next come the tasks to record patients' medical information and vital signs (45%) and to consult and coordinate with healthcare team members to assess (40%). Together they free the most hours across the industry.
How many hours can AI save a healthcare provider?
About 657 hours a year per full-time person, 33% of a 2,000-hour working year. For a team of 10 that is about 6,570 hours, the time of 3.3 full-time people. The figure is the share of task time AI can remove today, weighted by how many people do each job.
Which roles in a healthcare provider are most exposed to AI?
Of the roles with at least 2% of the jobs, medical secretaries score highest, at 65 of 100, and nursing assistants lowest, at 16. Each role's page lists its tasks one by one.
Will AI replace jobs in healthcare?
AI can do 14% of the task time across these roles today. Another 47% gets faster with AI while a person stays in charge. 39% needs a person. One task that stays with people is to administer medications to patients and monitor patients for reactions or side effects.
Where should a healthcare provider start with AI?
With the quick wins: approved assistants for the tasks AI helps with, under a written AI policy. Then automate the 5 tasks on this page AI can do end to end. An AI audit maps the tasks in your team and ranks them the same way this page does.
Next steps
Find the tasks to automate first in your team
We map the tasks in your business, score each one the way this page does and build the automations that pay back first.