AgenticMD is an AI clinical extender paired to each provider's panel. Patients text, the AI prepares a structured intake and draft order from Epic, and the provider approves, edits, or rejects in one tap.
Published time-motion and EHR event-log research shows how much of a primary care physician's day goes to the computer instead of the patient.
of an 11.4-hour workday spent in the EHR per weekday, per 1.0 clinical FTE (142 family physicians).[1]
spent on inbox management alone, 23.7% of total EHR time.[1]
EHR and desk work for every hour of direct patient care in ambulatory practice.[2]
Figures are from the cited studies. They are not AgenticMD results.
No app, no portal login. Native SMS for patients, with the provider in control of every clinical action.
Native SMS with talk-to-text on any mobile phone. Identity is verified before any clinical conversation.
A structured symptom assessment uses the patient's medications, allergies, and conditions from Epic (FHIR R4). A deterministic emergency screen runs first.
A clinical summary, with a draft order when appropriate, lands in the provider's queue to approve, modify, or reject. Nothing is autonomous.
These are calculator defaults, not measured results. Adjust every input to match your organization on the full calculator.
Defaults: 50 In-Basket messages/provider/day, 40% handled by AgenticMD, 2 min saved per message, $125/hr provider time, 240 workdays, 200 open care gaps/panel, 20% completion lift, $25/gap, $100/provider/month. Estimate, not a quote.
This is the reduction AgenticMD is being designed to reach. It has not been measured in a clinical setting. Measuring it against your baseline would be a primary goal of the founding-partner pilot.
Hard-coded pattern matching catches cardiac, stroke, respiratory, and other emergency categories before the AI ever replies.
Clinical content encrypted at rest, TLS in transit, PHI hashed in audit logs. Executed BAA with the AI model provider.
Every authentication, message, and provider action is logged. Zero autonomous clinical actions.
Read the full security & compliance status, including what is still pending →
AgenticMD has no customers, no pilot data, and no testimonials yet. The product is running pre-launch integration tests with synthetic, non-real patient data. We are looking for one healthcare organization to be our first partner and to help measure what it actually does.
No licensing fee for the pilot. In exchange, we ask partners to share de-identified, aggregate pilot outcome data, handled under a BAA, so we can measure results together.
Family medicine physician with ProMedica Physicians Group since 2016, practicing full-scope family medicine, including obstetrics and hospitalist care. Trained at the Texas A&M family medicine residency and Wayne State University School of Medicine.
AgenticMD is built around one rule: the AI prepares, and the provider decides.
“I’ve practiced full-scope family medicine since 2016, from delivering babies to rounding in the hospital. The care itself has never been what wears clinicians down. The inbox has. I built AgenticMD so the AI does the preparation and the provider makes the decision, every time.”
Joshua Lucas, MD, Founder
No. AgenticMD is pre-pilot and runs on synthetic test data only. We are looking for our first healthcare partner.
No. It is a design target. The founding-partner pilot is meant to measure the real effect against your own baseline.
No. Every clinical action requires provider approval, and emergencies are routed by a deterministic screen before the AI responds.
No licensing fee for the pilot. In exchange, we ask partners to share de-identified, aggregate pilot outcome data, handled under a BAA, so we can measure results together. Pilots run 3–6 months with 1–10 providers.
Epic, through FHIR R4 and SMART on FHIR.
Not yet. SOC 2 is on our roadmap and is labeled as planned. See Security for current status.
Book a 30-minute demo with the physician founder and see whether a founding-partner pilot fits your organization.
Book a 30-min demo