New York City-based Mount Sinai Health System’s pharmacy leaders are no longer debating whether to adopt AI. The question now is where to focus — and how to move from exploration to execution without introducing friction into daily workflows.
“We can all agree that we really would like to see some things happen this year,” John Manzo, PharmD, senior director of pharmacy services, said during a recent strategy discussion. But despite that urgency, the team is “still sort of in that exploration phase,” weighing which initiatives are ready to scale and which require further vetting.
That balance between momentum and caution is shaping how the health system structures its pharmacy AI strategy.
For Cathleen Mathew, PharmD, senior director of ambulatory pharmacy and coverage, the guiding principle is straightforward: Improve efficiency while ensuring pharmacists are practicing at the top of their license. In practice, that means evaluating AI across inpatient order verification, identifying high-risk patients during transitions of care, strengthening inventory management and refining prior authorization workflows.
Mount Sinai has launched AI-powered prior authorization and deployed ambient AI at the system level, with plans to expand it to ambulatory and transitions-of-care pharmacists. Even so, leaders continue to return to a central question: What problem are we solving?
“Is it actually folding into a workflow?” Dr. Mathew said. “We don’t want to make extra work; that defeats the purpose.”
Governance is critical to maintaining that discipline.
Pharmacy leaders help identify use cases and conduct early evaluations before proposals move through a formal pharmacy governance structure. From there, initiatives advance to broader system committees that include physicians, nurses and pharmacists.
That layered review prevents AI efforts from becoming siloed and ensures they align with enterprise priorities. Each proposal is examined from multiple angles, including operational feasibility, integration requirements and oversight considerations.
Technology decisions also hinge on interoperability. With Epic serving as the organization’s backbone, leaders prioritize solutions that integrate seamlessly within the EHR environment.
Beyond integration, the team is focused on measurable value. “What extra value does it bring us that we’re not already getting from one of our existing applications?” Dr. Manzo said.
Workforce strategy adds another dimension. While many health systems are turning to automation in response to staffing shortages, Mount Sinai’s pharmacy leaders report stable technician staffing, supported by a strong in-house training program. Still, they see AI as a way to rebalance and strengthen work.
One example under consideration involves automated dispensing cabinets. Instead of having technicians respond to refill requests throughout the day, AI could analyze usage patterns and determine more efficient refill timing.
The team is also working to systematize medication reconciliation at intake. “If it’s crappy when they come in, it’s going to be a challenge throughout,” Dr. Manzo said, referring to inaccurate medication lists at admission. Pharmacy interns and technicians already support that process, and leaders see potential to embed AI into that workflow.
As initiatives move forward, evaluation remains practical and hands-on. Leaders assess whether a solution reduces manual steps, integrates cleanly with Epic and fits into daily operations. Early feedback on the AI-powered prior authorization tool has been “strong and positive,” Dr. Mathew said — an encouraging signal as additional projects are considered.
Looking ahead three to five years, pharmacy leaders envision deeper AI integration across medication management, supply chain operations and patient communication, including conversational or agentic tools to support medication education and outreach.
For now, the priority is shifting from assessment to action. The remaining hurdles are practical: deciding “what and who we’re going to partner with,” allowing time for build and testing and aligning investments across governance layers.
