How Genentech extracted information for clinical trial analytics with Snorkel Flow

Genentech, part of the Roche Group, manages thousands of Clinical Trial Protocol PDFs a year and needed faster access to trial objectives, methodology and population data to cut development costs and timelines. Using Snorkel Flow, Genentech built an inclusion/exclusion criteria pipeline (NER, entity linking, rules-based relationship extraction) identifying 21 CMS Chronic Condition Entities at 98% accuracy in a few weeks, plus a Schedule-of-Assessment pipeline classifying extracted procedures into 8 categories. Overall model accuracy reached 95-99%, replacing months of manual labeling; Principal Data Scientist Michael D'Andrea said "Snorkel Flow made the pipeline development a drag and drop experience."

Snorkel AI is a data-development and evaluation company whose programmatic labeling tools build training and benchmark datasets used to stress-test whether a model approach is viable before a full build.

This sits in the Design & Stress-Test stage of the AI use-case lifecycle: Named examples from the tools pharma teams use to scope, prototype and benchmark an AI use case before committing to build it — data-catalog, sandbox and evaluation platforms.