The specialty drug pipeline is outrunning the infrastructure built to get those drugs to patients. Latent, an enterprise pharmacy intelligence platform, is betting that autonomous AI agents — not tools that merely flag or recommend — can close that gap for a roster of health system partners that’s grown to 60, with seven of the country’s largest systems announced Sept. 15.
Cleveland Clinic, AdventHealth, Cincinnati Children’s, Keck Medicine of USC, UNC Health, UnityPoint Health and WVU Medicine are going live on the Latent platform, bringing the company’s footprint to 60 health systems serving an estimated 70 million patients combined.
Latent also unveiled a suite of pharmacy agents built to mine patient charts, claims, payer policies and drug labels to work through prior authorizations, compliance reviews and work queues for high-demand drugs, including GLP-1s and specialty therapies in immunology, rheumatology and cardiology. The announcements arrive alongside a refreshed brand and new website for the company, which is on pace to process more than 5.8 million patient medication cases over the next 12 months.
The moves come as pharmacy operations face continued pressure. Active U.S. drug shortages continue to climb as specialty drugs are projected to grow from $225 billion in 2024 to $2.3 trillion by 2034. That pressure compounds a workforce already stretched thin. Becker’s reporting found the same warning from pharmacy leaders nationwide: persistent shortages of pharmacists and technicians, worsened by burnout. Latent, founded in 2022, is betting that autonomous agents, not incremental fixes, change the math.
“Advancements in medicine are outpacing the infrastructure built to get new therapies to patients,” said Sri Somasundaram, co-founder of Latent. “A treatment can exist, work, and still never reach the person who needs it because access is mired by complex coverage rules. We’re giving health systems the ability to extend the capacity of their pharmacy teams and get patients into the right treatment, without an unnecessary wait.”
‘The tipping point’
When Medicare’s GLP-1 Bridge Program took effect July 1, 4 million Medicare members became eligible for coverage overnight. Latent deployed agents within 10 days to autonomously evaluate eligibility and gather the information needed to submit prior authorizations — compressing a process that would typically take health system pharmacy teams far longer to stand up manually.
“The tipping point for us was recognizing that technology had matured enough to help prioritize pharmacist attention,” Charles Barkey, vice president, information technology of Northern Panhandle WVU Medicine Hospitals, told Becker’s of the system’s new partnership. “We weren’t looking to replace clinical judgment. We were looking for a way to help our pharmacists focus on the interventions that truly impact patient outcomes.”
Mr. Barkey noted that WVU Medicine has already seen efficiency gains. With less time spent on routine order verification, pharmacists have more room for complex medication management, direct patient care, provider consultation and interdisciplinary collaboration, while patients get faster decisions on the approvals they’re waiting on.
Robert Granko, Pharm.D., system vice president of pharmacy business operations at UNC Health, said intelligent automation is already changing how his team manages workload. “We have seen how intelligent automation can help our pharmacy teams manage growing demand, improve visibility across workflows, and accelerate access to therapy,” Dr. Granko said, adding that it frees staff to focus “where it matters most” and get patients started on treatment sooner.
What it takes to earn a health system’s trust
Latent’s earliest customers, Yale New Haven Health and Ochsner Health, were themselves large, complex systems — so the addition of Cleveland Clinic, AdventHealth, WVU Medicine, UNC Health and others hasn’t changed Latent’s core challenge so much as intensified it, Mr. Somasundaram told Becker’s.
The biggest obstacle to adoption hasn’t been technical performance, he said — it’s fear, and it takes two forms: whether the AI can be trusted to get the answer right, and what happens to pharmacy staff if it does.
On the first, Mr. Somasundaram pointed to outcomes as the answer. “We have been maniacally focused on getting these agents to perform incredibly well in their core function of facilitating that dumpster diving of gathering the patient data to answer a question and substantiate whatever workflow we’re supporting in pharmacy,” he said. “The proof is in the pudding.”
On the second, he said Latent leads with health system leadership from day one to frame automation as capacity, not replacement. “There’s only a growing patient population. There’s only demand to help grow the program. And so it’s not about really taking away roles,” he said. “It’s actually just about how do we continue to handle more of this volume, continue to handle more patient demand, and give them the tools to succeed so they’re not buried under that work.”
Mr. Barkey said he expects AI’s role in pharmacy to keep expanding in the years ahead, becoming “a more trusted operational and clinical assistant that helps pharmacists and support teams prioritize work, identify risks earlier, and make more informed decisions across the medication-use process.”
