Pharmacy and clinical teams are operating in one of the most demanding environments in recent memory. Staffing shortages, accelerated by the pandemic and compounded by an aging patient population, are forcing health systems to do more with less — and pharmacy leaders are sounding the alarm on a brewing workforce crisis. Cost pressure has only sharpened, with tariffs pushing pharmacy executives to redo budgets frequently and patients with multiple chronic conditions driving new complexity into drug spend.
In this environment, AI is being layered — fast — into workflows that touch patient care most directly. Medication management is one of the highest-stakes places it can land. And medication intelligence, the layer above raw drug data that synthesizes, combines and points clinicians to the next best action, has become foundational to whether AI delivers reliable, scalable, real-world impact.
Yet despite the dizzying pace of innovation, two things remain core to pharmacy operations: trust and an unrelenting focus on reliability.
“There isn’t a lot that’s actually changing. That core principle of being right is still fundamental,” Staci Hermann, PharmD, vice president of Medi-Span and payer strategies at Wolters Kluwer Health, said.
What is changing is the bar for the AI tools clinical leaders bring in. To meet it, Dr. Hermann points to three strategies to ensure advancements in medication intelligence remain trusted, scalable and designed to focus on outcomes.
AI tools must be clinician-led, starting from development
The volume of AI tools pitched to healthcare executives is staggering, with many systems scrambling to scale responsibly and close governance gaps. But not all AI is created equal; and in clinical pharmacy, the difference shows up fast.
“Bad input data leads to bad output data, and the same is true with AI,” Dr. Hermann said. If the foundational drug content is inaccurate, she explained, AI’s recommendations will be too. That makes two things non-negotiable: traceability and the people building the system.
Health systems should be able to see where an AI agent retrieved its information, understand who is curating the underlying drug content and know who authored the prompts shaping the model’s responses.
At Wolters Kluwer, the editorial team curating UpToDate and Medi-Span content is staffed by clinicians — many still practicing — and those same subject matter experts develop the prompts behind the AI. “To me, you need both, because at the very end of this whole process is a patient,” Dr. Hermann said.
That principle sits underneath Wolters Kluwer’s recent unveiling of Medi-Span Expert AI, beginning with a Model Context Protocol (MCP) server that connects expert-curated medication data to digital health workflows. It follows UpToDate Expert AI, launched in 2025. Dr. Hermann is clear that it is still early days, but she expects gains in patient care, cost management and throughput as the tool matures — mirroring the documented impact of the clinical decision support tools that came before it.
Question AI tools ruthlessly: 5 key considerations
Healthcare leaders are scrutinizing AI more than ever. When it comes to medication intelligence and adopting new tools, Dr. Hermann urged leaders to consider five questions:
- What is the clinical foundation the tool is built on, and who is defining the logic?
- How explainable are the outputs? Are they predictable, sourced appropriately and reliable?
- Can you audit the AI?
- What clinical governance does the developer have in place internally?
- Is it a vendor that has been providing clinical content historically, or a newer entrant — and how does that shift your risk tolerance?
Just as important: do not rush the pilot phase. When assessing new AI tools, Dr. Hermann recommended that pharmacy and clinical teams pull a complicated patient case and run it through the system to see whether the AI arrives at the same conclusion the clinician does independently. Edge cases like rare diseases or atypical doses are where hallucinations surface, and they warrant the most rigorous testing, she said.
Make trust the north star for partners and internal teams alike
Even with rigorous vetting up front, the trust health systems place in their AI vendors is fragile — one stretch of inaccurate outputs can unwind it. “You start to lose trust in that product, in that vendor,” Dr. Hermann. “As a health system, I would not want to use it because I just don’t trust what it’s going to tell me.”
Internal trust matters just as much. The organizations doing this best, Dr. Hermann said, share a close partnership between clinicians and IT or informatics teams — a contrast to the all-too-common pattern of siloed IT decisions that leave clinicians using tools they were never consulted on.
Three factors tend to drive sustained adoption: reliability and predictability, so clinicians get consistent answers in the same clinical context; clear alignment around patient safety; and a deliberate framing of the technology as augmentation, not replacement. Tools that fit naturally into existing workflows, rather than asking clinicians to log into yet another system, earn that trust fastest.
Looking ahead: rising complexity, sharper partnerships
Over the next few years, Dr. Hermann is closely watching two converging forces — a robust pipeline of new medications and an aging population that will place unprecedented demand on health systems.
“We are entering an even more complicated state for healthcare going forward,” she said. “Having the ability to go through and process all different types of information — whether it’s questions, whether it’s clinical or financial — those become core fundamentals to hospitals staying open, to patients getting on therapies when they need them, and to patients being able to afford them.”
Her closing advice draws on her experience as a former chief pharmacy officer: finding a trusted vendor, and partnering with them deeply. “If I knew that vendor was already in a couple of other health systems I admired or were close to in size, it helped me go, ‘Okay, there’s something there I want to explore.'”
In a market crowded with AI promises, that signal — paired with clinician-led content and a demonstrated focus on trustworthiness — is what separates expert AI from the rest.
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.
