🎧 LISTEN | On the go? Start here.

Listen on Spotify or Apple Podcasts.

Why 90% of Trials Still Fail and What AI Can (and Can't) Fix

A trial can fail for reasons well beyond the molecule: population, indication, dose, design, sites, monitoring, and data interpretation all shape what a study reveals. On AI for Pharma Growth, I joined Dr. Andrée Bates to discuss where AI can improve these decisions and where it can't replace scientific and clinical judgment. We cover how causal machine learning, adaptive design, and large-scale simulation help sponsors identify responder subgroups, test designs before enrollment, and connect protocol design with clinical operations. One theme runs through it all: a confident prediction isn't enough. Sponsors and regulators need to see the data, reasoning, and evidence behind a recommendation.