The workup starts before the doctor walks in.
Laminar MD builds AI-EDdoc, a decision engine that works inside the EHR. It drafts the workup, pends the highest-yield orders, and keeps the can't-miss diagnoses on the table, so emergency physicians move faster without cutting corners. The physician signs every order.
Emergency medicine runs on speed and certainty. Today it gets neither reliably.
Sources: CDC National Center for Health Statistics, NHAMCS 2022 · Agency for Healthcare Research and Quality, Diagnostic Errors in the Emergency Department (2022) · Mello et al., Health Affairs (2010), 2008 dollars · Physician-in-triage estimate: $3.4M/yr at one 400-bed hospital (staffing schedule × ~$300/hr loaded physician cost) × ~5,600 U.S. EDs
Missed diagnoses
The dangerous diagnosis rarely announces itself. Atypical presentations and crowded shifts are where misses happen, and diagnostic error drives patient harm and malpractice exposure.
Defensive overtesting
Without a reliable safety net, clinicians order broadly to protect themselves. Patients wait longer, departments clog, and costs climb.
Triage that can't scale
Putting a physician in triage gets the workup started sooner, but few departments can staff it every hour. AI-EDdoc is designed to start the right workup at triage without that extra shift.
Five steps, inside the workflow clinicians already use.
No new screen to learn. AI-EDdoc runs in the background and lands its work in the EHR as a draft note and pended orders.
- Intake Multilingual triage
- Reason AI engine + guardrails
- Pend Draft note + orders
- Sign Physician decides
- Re-assess Re-reasons as results return, through to disposition
Illustrative workflow. Demonstrations use synthetic patient data only.
Learned from outcomes, not opinions.
Grounded in real encounters
Our engine is built on hundreds of thousands of real-world emergency visits, working backward from final diagnoses to the orders that actually found them.
Speed and safety together
It front-loads the highest-yield workup and adds a safety net for the can't-miss diagnoses that hide behind atypical presentations.
Lives in the EHR
Standards-based integration (SMART on FHIR, HL7). Recommendations arrive where clinicians already work.
Gets sharper every month
Every recommendation is audited against what physicians actually ordered and how the visit ended. Updates are physician-reviewed before release.
A safety spine, not a black box
- Nothing goes live without a physician's signature. AI-EDdoc proposes; clinicians decide.
- Every answer passes can't-miss coverage, allergy, renal, and duplicate-order checks.
- Recommendations come with evidence-based rationale clinicians can verify.
- Protected health information stays within the EHR and BAA-covered infrastructure.
Past the slide deck.
Built in Epic
Our Epic integration is built and registered, with orders pended for physician signature inside the EHR.
First deployment underway
We're working with a community health system as our design partner and first deployment site.
Why now
Federal interoperability rules opened EHR data through FHIR, and modern AI can now reason over it. The missing piece is clinical judgment built in. That's us.
Built in the emergency department, for the emergency department.
Ami Kurzweil, MD
Practicing emergency physician, ultrasound fellowship-trained, and educator of emergency medicine residents. Previously Assistant Director of Operations at SonoSim, helping scale a medical-education technology company. Ami builds AI-EDdoc from the bedside: every rule in the engine answers to a clinician.
Bring AI-EDdoc to your emergency department.
We're partnering with a small number of health systems now. Clinical leaders, CMIOs, and investors are welcome to reach out.
AI-EDdoc is clinical decision support; all clinical decisions are made by licensed clinicians.
© 2026 Laminar MD Inc., a Delaware corporation.