AI Pharmacist

Problem

In hospital pharmacy workflows, patients, clinicians, and pharmacists repeatedly ask about medication use, prescription appropriateness, and polypharmacy risk. These questions require checking labeling, guidelines, interaction rules, and dose adjustments for special populations. If an AI answers without constraints, it can overstep into diagnosis, give unsafe dosing, or blur professional responsibility. AI Pharmacist frames pharmacy Q&A as an explainable, traceable assistance process rather than a replacement for pharmacists.

How It Works

The skill first performs Phase 1 role, intent, and urgency identification, confirming whether the user is a patient, pharmacist, manager, or clinician, then classifies the request as routine, monitored, or urgent. It searches the knowledge base, generates a structured response, and runs dual review. Core modules include:
- Medication queries: usage, adverse reactions, and adjustments for pregnancy, lactation, elderly, or hepatic/renal impairment;
- Interaction checks: risk levels such as contraindicated, use with caution, monitor, and no clinical significance;
- Prescription review support: red, orange, and yellow alerts for contraindications, excessive doses, duplication, or inappropriate routes;
- Pharmacy education: Pharm-E prompt framework, case discussion, and AI-assisted review skills.

Responses lead with the core conclusion, then add analysis, source notes, and risk disclaimers.

Boundaries

It is useful for drug information, interaction assessment, prescription review assistance, smart pharmacy service design, and pharmacy teaching. It is not intended for clinical diagnosis, surgical decisions, emergency medication decisions, or final pharmacist review responsibility. Listed references/ and scripts/ can be loaded or executed only if the deployment provides them and validates the underlying data.