In this video interview, Patrick Mizer, chief technology officer at Ledger Run, addresses the reality that 80% of sites operate on six months or less of cash—and explains why growing investment in agentic AI and workflow modernization signals that the industry is starting to treat payment infrastructure as the priority it deserves.
In a recent video interview with Applied Clinical Trials, Patrick Mizer, chief technology officer at Ledger Run, discussed how payment reliability has quietly become one of the most consequential competitive differentiators in clinical research—and why the infrastructure behind site payments is overdue for a fundamental redesign. He opened by reframing the competitive dynamic in clinical research: as experienced sites gain more study opportunities than they can support, sponsors are increasingly competing not just for patients but for the attention of high-performing investigative sites. In that environment, payment performance has become an operational brand—and sponsors with a reputation for predictable, accurate payments are gaining an advantage in site selection.
Mizer traced the persistence of payment inconsistency to decades of incremental growth without intentional design. Study budgets, contracts, CTMS platforms, EDC systems, and financial systems have remained disconnected, each introducing its own manual processes and assumptions. The operational pain, he noted, has historically been distributed widely enough across sponsors, CROs, and sites that no single party felt compelled to fix it wholesale.
On AI, Mizer was measured but specific. Skeptical of broad claims about AI's transformative potential—he invoked the pets.com era as an analogy—he described where AI is genuinely effective today as low-level reasoning tasks that are labor intensive, have QC built in, and don't yet require true expert judgment. Invoice processing is Ledger Run's proof point: 60% of customer time was being spent reading, translating, and manually entering invoices. The company built an agentic AI flow that ingests emailed invoices, screens for PHI, decomposes the components, and validates them against the system—with accuracy high enough that QC experts could not distinguish AI outputs from human work in blind testing.
He closed by connecting payment infrastructure directly to the broader site access problem, noting that with 80% of sites operating on six months or less of cash, the urgency is clear—and that the industry's growing investment in workflow modernization and agentic AI is a sign that the problem is finally being treated with the priority it deserves.
