Case Study: Top 3 US Bank achieves 2,000+ hours/year savings for financial analysts with Snorkel AI

Top 3 US Bank - Customer Case Study

Top 3 US Bank analysts were spending hundreds of hours a year manually reviewing 10‑K financial documents to find information on interest rate swaps, which limited their ability to assist customers proactively. The team saw the potential of AI/NLP but lacked the labeled training data needed to build a reliable extractor, and engaged Snorkel AI using its Snorkel Flow platform to address the gap.

Snorkel AI applied programmatic labeling and weak supervision—encoding analyst expertise as labeling functions—to train a custom NLP model that automatically identifies and extracts interest rate swaps, reaching an F1 score of 83. The solution programmatically generated 70k labels per minute via Snorkel Flow, saved 2,000+ analyst hours per year, and produced a production-quality AI application in six weeks.