Practice and Tools
Evaluate pharmacy technology and AI against the work it must support
- Summary
- A compelling demonstration is a starting point. A useful evaluation examines the intended use, the evidence, the workflow, and the controls needed in the actual setting.
- Last reviewed
- September 29, 2026
How should a pharmacy evaluate a technology or AI proposal?
Evaluate technology in the workflow where it will actually be used. Intended use, evidence, information handling, human oversight, failure modes, integration, and monitoring all matter before a tool is described as safe or effective in practice.
Pharmacy professionals bring the medication-use and safety expertise that technology decisions need. Begin with the intended use and the decision the technology would support. This PharmD Hub™ checklist helps pharmacy teams and leaders examine evidence, information handling, human responsibilities, and failure planning before adoption. It supports an organizational review; it does not certify a product or establish regulatory compliance.
Define the Intended Use
State the problem, users, inputs, outputs, and decisions the technology would support. Identify what it is not authorized to do. Compare the proposal with the current process and simpler alternatives.
Examine Evidence and Limitations
Ask how performance was evaluated, in which population or setting, and under what conditions. Request examples of failure, uncertainty, and limitations. For AI generated content, require verification of references and clinically relevant assertions; a plausible response is not proof of accuracy.
Review Information Handling
Use your organization’s approved security, privacy, procurement, and legal processes. Confirm what data are collected, retained, accessed, or used for other purposes. Do not enter patient or confidential information into an unapproved service.
Plan the Human Workflow
Identify who reviews outputs, who may act on them, and who is accountable for decisions. Consider training, accessibility, alert burden, documentation, and the possibility that users will rely on outputs beyond the intended purpose.
Pilot decision prompts: What must the system demonstrate? Which errors would make use unacceptable? Who can pause it? What would trigger additional training, restriction, or withdrawal? Agree on the criteria before judging the pilot’s results.
Plan for Failure and Change
Define escalation, downtime procedures, monitoring, and reevaluation after material changes. A successful pilot does not eliminate the need to review performance in routine use.
Scope note: This is an organizational evaluation framework, not a technical certification or a determination of regulatory compliance.
Bring the Checklist to Your Next Review
Draft your questions before a vendor demonstration and record which answers require evidence or specialist review. Revisit them after testing in the intended workflow. The service proposal framework helps connect the technology decision with the wider service need.
Next in Practice and Tools
Clinical Calculators