Paperwork, rotas and intake forms come first. Clinical calls stay with clinicians.
81% of US physicians now use AI, and the top uses are summaries and paperwork. Learn to time a real task, size the check to the stakes, and keep patient data out of public tools.
The AMA's 2026 physician survey found that 81% of doctors now use AI professionally, more than double the 2023 rate. The top uses are paperwork: summarizing research and care standards (39%), drafting discharge instructions, care plans or progress notes (30%), documenting billing codes and visit notes (28%), and drafting replies to patient portal messages (19%). Assistive diagnosis sits at 17%. Asked what is critical to wider adoption, 88% named safety and efficacy validation and 86% named data privacy.
That pattern matches what a charge nurse, a scheduler or an intake coordinator sees on the floor. The tools arrive in the inbox and the EHR first. Clinical calls stay where they were.
About 24 million people work in health care and social assistance, and BLS projects about 1.9 million openings a year in healthcare occupations this decade. Most of those people spend part of every shift on text and tables rather than patients.
Privacy first. Under HIPAA, a vendor that creates, receives, maintains or transmits protected health information on behalf of a covered entity, such as a practice or hospital, is a business associate. The covered entity needs a written business associate agreement with that vendor before sharing the data. A public chat assistant on a personal account is not that vendor. Nothing with a name, a date of birth, a record number or a face goes into it. Use the tools your organization has contracted, and ask the privacy officer which ones those are.
Accuracy second. The Start Here course has you size the check to the stakes, against a source outside the chat. For a rota email, a glance at the schedule. For a dose, the drug reference and your workplace's second check, every time. In the course's worked example, a charge nurse asks an assistant for a pediatric dose and the ward's drug reference gives a different one. She uses the reference. The course also names four reasons an answer goes wrong, including invented citations, which matters when an assistant summarizes a care standard that may not say what it claims.
Clinical decisions stay with clinicians. Nothing on this page, or in these courses, is diagnosis or treatment advice.
Charge nurses and unit clerks, medical records specialists (BLS projects 8% growth for that role from 2025 to 2035), schedulers, practice managers, intake and referral coordinators, billing and coding staff, and clinic administrators.
Start with the free Start Here course. It takes about two hours, needs no card, and follows a fictional charge nurse: timing her weekly timesheet chase by hand and with an assistant, then sizing the check on a dose question and a rota email. You end with a one-page verdict memo on a task of your own.
If your work is admin flow, such as intake forms, referral routing or records requests, the automation course has you cost a manual task from timed runs, build one form-to-CRM intake on Zapier, Make and n8n, check the record each one wrote, route each item to one owner with a fallback, and add one fenced AI step with a daily spending cap. Build it on test data, and bring the privacy officer in before any real patient field passes through a third-party tool.