Finalist
Engineering Trust: A Compliance First AI Engine for Medical Legal Review
Summary of work
Medical Legal Review (MLR) is one of the most time consuming stages of medical communications. Multiple review cycles, repeated amends and manual re‑checks place a burden on clients and agencies, slowing delivery and absorbing time in work that is often mechanical rather than scientific.
Together with Ipsen, Syneos Health explored whether AI could responsibly reduce this burden at the most sensitive point: before MLR begins. Rather than pursuing automation, the proof of concept was designed to remove only clear validation tasks, while preserving interpretive judgement for human reviewers.
The resulting AI‑assisted reference validation system extracts citation‑linked claims with 100% accuracy, retrieves evidence exclusively from approved libraries, and produces structured, audit‑ready substantiation reports in approximately 40 minutes — a task that typically requires four to ten hours manually.
Pilot testing demonstrated 85–91% alignment between AI outputs and expert reviewer assessments. In more ambiguous cases, where professional judgement plays a greater role, the system consistently took a more conservative position. In 100% of these instances, reviewers agreed with the justification for caution, confirming the system’s value for trusted support. By handling structured validation and surfacing uncertainty early, the AI enables reviewers to focus their expertise where it matters most, without relinquishing control.
Judges’ comments
The metrics were very impressive in this entry from Syneos Health Communications and Ipsen. It had strong execution and the judges liked the interview-led process. Exciting to see what they’ve done.
