Although artificial intelligence has many benefits, the technology is not appropriate for everything; human expertise will always be needed in pharmacy, noted leaders from CaryHealth, a Washington, D.C.–based provider of AI-powered tools for healthcare.
CaryHealth leaders teased out when to use AI—and when not to—during a session at Asembia’s AXS25 Summit, in Las Vegas.
“AI is good at making the patient experience better by freeing up pharmacists to provide more patient care,” said Matt Hawkins, BSC, CaryHealth’s chief technology officer. The technology also excels at assisting with inventory management and drafting patient communications, both of which save pharmacist time. AI agents also can expedite transferring a patient to a live pharmacist when medical advice is needed, by triaging these calls more quickly than possible with human-to-human relays.
AI can also help pharmacists make better clinical decisions, noted CaryHealth CEO Areo Nazari, PharmD. “Large language models and AI can help answer complex clinical questions,” he said, giving this example: What’s the optimum dose of IV vancomycin for a 55 kg, 70-year-old man with poor renal function? Rather than poring through textbooks or pecking through PDFs, using an AI-powered clinical decision support tool can yield an answer to this question much faster. “It will give you a very precise IV dosing regimen and also give you things to look out for,” Dr. Nazari said, with references included.
But there are limits to what AI can do, as well as tasks that AI should not be given—even if the technology allows it, Mr. Hawkins stressed. For example, “you should not use AI to put barriers between a patient and a pharmacist,” he said, by requiring extensive engagement with a bot before someone can finally speak to a human. Instead, the AI bot can find a time for the patient and pharmacist to meet and schedule that appointment. “That’s an efficient use of AI.”
AI also should not be used to deliver someone’s lab results or to explain a complex diagnosis with potentially difficult treatment options. AI can filter this kind of information for pharmacist review, but the pharmacist should always deliver the results to the patient, from one human to another, Dr. Nazari noted, adding, “AI should enable our pharmacists to be incredibly human.”
The sources reported no relevant financial disclosures.
This article is from the August 2025 print issue.