Redbud Hollow Medical Group (a fictional composite: 16 providers, two sites near San Antonio, 18.4% of calls abandoned, two scheduler seats empty and a partners' budget meeting on December 10) runs through every module. You decide what staff may paste into which assistant after a survey finds patient details in personal chatbots and two clinicians recording visits with a consumer app; build the practice's AI inventory, including features a vendor turned on by default; compare an EHR add-on phone agent, a stand-alone agent, two more schedulers and an overflow service, then write the routing, red-flag escalation, caller-identity and disclosure rules and audit the agent weekly; check outbound AI-voice and text scripts against the TCPA health-care exemption; test a no-show model by patient group before acting on it; set auto-file rules for document AI and handle a misfiled lab result; check an AI-drafted self-pay Good Faith Estimate; run an ambient-scribe pilot with a comparison group, a consent workflow and a note audit, and write the keep, expand or stop memo; reconcile an AI-built denial analysis to the source system and draft appeals in which every sentence traces to the record; and report it all to the owners with measured results separated from projections and cash separated from capacity. Each module ends in a Decision Lab where the fast option is tempting and its cost arrives later, a 30-minute capstone simulation runs the whole season to the partners' meeting, and the capstone project combines your deliverables into a Practice AI Operating Plan, offered as practice with a self-scored rubric.
Real products (Microsoft Copilot Chat, Claude Enterprise, ChatGPT for Healthcare, eClinicalWorks, athenahealth, Assort Health, Phreesia, Luma Health, Tennr, Dragon Copilot, Abridge) are compared only by what their makers publish; vendor performance figures are treated as claims, never as evidence. The rules are taught where they bite: HIPAA minimum necessary, business associates, Safe Harbor de-identification and breach risk assessment and notice (45 CFR 164.502, 164.514, 164.402, 164.404, 164.408) and Texas's own breach-notice duty to the attorney general (Tex. Bus. & Com. Code 521.053); FCC 24-17 and 47 CFR 64.1200(a)(9)(iv) and (a)(10); Texas AI disclosure (Tex. Bus. & Com. Code 552.051) and Texas SB 1188; Section 1557 at 45 CFR 92.201 (language access) and 92.210; Good Faith Estimates at 45 CFR 149.610 and 149.620 (and HHS's enforcement discretion for co-providers); the information-blocking rule at 45 CFR Part 171; the False Claims Act and CMS-0057-F denial reasons. Coding is covered in hls-155, building claims and prior-authorization automation in hls-255, and PM/EHR integration in hls-275. Status is stated as of October 2026; the course is educational, not legal advice, no vendor endorses it, and every patient, call, fax, note and remit in it is fabricated.