Steve
Herne

Somewhere around year fifteen of watching pharmaceutical companies take an average of twelve years to bring a single drug to market, Steve Herne started asking a different question. Not "how do we run this trial faster?" but "what if we didn't need as many people in the control group to begin with?" That question - which most of the industry had decided was rhetorical - turned out to have an answer. He's now running the company that built it.

Herne's path to Unlearn.AI is a career study in how domain expertise accumulates into something rare. He started in pharma services at Inveresk Research in Scotland, then worked his way through MDS Pharma Services, Covance, Icon Development Solutions, ERT, Bioclinica, and WCG - each move sharpening his understanding of what drug developers actually needed versus what the industry was selling them. At Bioclinica and WCG, he held the Chief Commercial Officer role, which gave him a front-row seat to every friction point in clinical development. The bottlenecks were not mysterious. They were structural.

Unlearn.AI was founded by mathematician Charles Fisher with a specific bet: that generative AI could create individual "digital twins" for clinical trial participants - virtual models that predict how each patient would have responded to placebo, even if they never actually received it. If the model is accurate enough, trial designers can shrink the placebo group significantly without sacrificing statistical power. Fewer patients take a sugar pill. Trials cost less. Results arrive faster. The science took years to validate. The European Medicines Agency qualified the methodology. The FDA issued guidance aligned with the approach. By early 2024, Unlearn had raised $130M and was looking for someone who could turn a validated research platform into a commercial business.

That someone was Herne. He joined in May 2024 as the company's first Chief Commercial Officer - specifically hired to do what the founding team, deep in AI and computational biology, hadn't fully done: build the go-to-market machine. Four months later, in September 2024, he was named CEO. The transition was characteristically unflashy. Herne describes it simply: "It's been an excellent transition, and I credit that to Unlearn's people and culture."

The challenge Herne walked into was a familiar one in deep-tech: a technically superior product that customers hadn't yet made habit. The pharmaceutical industry, as he puts it, are "laggards on the adoption curve" - companies that have run trials the same way for decades and treat every new technology as a risk to manage rather than an advantage to capture. His response was not to harangue them but to challenge his own team. Most of Unlearn's engineers and data scientists had never worked on a clinical trial. Herne started requiring them to "step into our customers' shoes" - to understand the operational pressure of a trial that's running behind, the regulatory risk of a failed submission, the human cost of asking 400 patients to sit in a placebo arm when you might only need 250 of them.

The results of 2024 made the case for the approach. AbbVie used Unlearn's digital twin models and presented retrospective data at the AD/PD Vienna conference showing a 23% reduction in placebo arm size in Alzheimer's disease studies. Unlearn's own research demonstrated a 38% reduction in control arm size for a Parkinson's disease endpoint - numbers that would have seemed implausible from a startup five years ago. The company reached record customer numbers, with sponsors returning for expanded programs across additional therapeutic areas. Herne is now planning an oncology expansion for 2026.

He is based in Philadelphia, Pennsylvania - not unusual for a pharma industry veteran, a world that runs on relationships built in the corridors of the industry's East Coast hub. He has a bachelor's degree in Pharmacy Instrumentation from Glasgow Caledonian University and a master's degree from the University of Strathclyde, the Scottish roots that sit quietly behind decades of American pharma career. He rarely discusses them. He rarely discusses much that isn't directly related to the mission, which he states with the clean simplicity of someone who has stopped bothering with the hedge: eliminate trial and error in medicine.

The pivot he is executing at Unlearn - from research-driven to commercially-executed, from bespoke consulting engagements to a scalable product platform - is the kind of work that rarely gets headlines but determines whether AI companies in life sciences become tools that researchers actually use. His 2025 goal, stated plainly: have Unlearn written into trial protocols before drug development begins, not called in as a late addition. When you're writing the protocol requirements before the sponsor files their IND, you are no longer a vendor. You are part of the infrastructure.