The safety check
Flags the discrepancy. A pharmacist resolves it.
Human-approved · flag-gated per tenantA real day with it
At discharge, the sentinel flags that the cardiologist's new anticoagulant duplicates an admission med under a different brand.
The pharmacist resolves it before the family is at the counter.
- 1
A patient is admitted or discharged with a medication list to reconcile.
- 2
The agent checks it against RxNorm and openFDA and flags discrepancies.
- 3
A pharmacist reviews and resolves every flag before it closes.
- 4
A dangerous duplicate is caught before the family reaches the counter.
Pharmacist
A review queue with findings, not a guessing game.
IPD doctor
Reconciliation status on the discharge checklist.
- Flag
- aiagents.medrec.enabled
- Reads
- Admission and discharge medication lists
- Never writes
- Never changes a dose — the pharmacist resolves every flag
- Channels
- Pharmacist review queue, checked against RxNorm and openFDA
- Languages
- English
Questions about this agent
Can it change a patient's medication?
No. It flags discrepancies against RxNorm and openFDA; a pharmacist reviews and resolves every flag before anything changes.
Is the AI the first check or the last?
Deterministic rules run first. The AI narrative is there to help the pharmacist understand a flagged discrepancy, not to replace the rule engine.
When does it run?
At every admission and discharge, wherever a medication list needs reconciling.