Case Study: a big four consulting firm eliminates a year of labeling time with Snorkel AI

How a big four consulting firm used NLP to monitor news for audits with Snorkel Flow

A big four consulting firm wanted to use AI to help its global auditing teams quickly find relevant news, regulations, and risk signals in dense unstructured documents, but it struggled to create enough high-quality training data for its machine learning models. The manual labeling process was slow, expensive, and hard to adapt as business needs changed.

Snorkel AI helped the firm build an NLP application using its programmatic labeling technology and annotation suite, allowing the team to label more than 10,000 news articles in a few hours instead of nearly a year of manual work. Snorkel AI also improved model quality, lifting F1 from 70 to 85 and increasing accuracy by 15% over hand labeling, while making development 3x faster and reducing SME time by 50%.