Global bank saves 10,000 hours in KYC efforts using Snorkel AI

A global custodial bank's analysts spent over 10,000 hours a year manually reviewing and transcribing 10-K filings for KYC verification (30-90 minutes per document), constrained by rigid rule-based systems and the need for scarce SME time to label data. Snorkel's team built an extraction application pulling 50+ attributes - nature of business, location, senior management - from tables, raw text and multi-page PDFs, aggregating results into structured data. The solution cut per-document assessment time from up to 90 minutes to 1-3 seconds, saved 10,000+ analyst hours, and achieved an +86 F1 macro score for risk profiling using just 25 hours of SME time, built in a few weeks.

The customer is anonymised in Snorkel AI’s own case study — not withheld by us.

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.