Impact
30x
Faster model delivery
>20
Point accuracy lift
67%
Increase in throughput
The challenge
The challenge
A leading global firm transforming insurance subrogation operations with AI found that manual review processes capped their throughput to ~30% of available claims. This bottleneck left significant revenue on the table and froze their ability to scale. The path to automation was further blocked by severe data imbalances where the critical signals for coverage appeared in only a small fraction of claims, making traditional AI models unreliable.
The solution
The solution
The firm engaged Snorkel to co-develop a comprehensive data-centric solution. By combining Snorkel’s team and AI research expertise with internal expert and data scientist collaboration, the firm was able to utilize programmatic data labeling and data-centric error analysis to rapidly correct training data at scale. This foundation enabled the development of distilled specialist models, turning a fragmented manual process into a high-accuracy automated workflow optimized for production.
The outcome
The outcome
Within 60 days, the team deployed a production-grade model that removed the manual bottleneck. Using a research-led approach the firm improved model performance by more than 20 points and accelerated the time to deliver new models by 3x. This efficiency expanded processing capacity by 67% (from 3,000 to all 5,000 claims), unlocking additional revenue from claims that were unable to be reviewed through manual efforts.
More customer stories
View all stories
DIU enhances decision-making resilience with Snorkel AI
Strategic dominance in the Indo-Pacific relies on the ability to track and coordinate friendly forces — ”blue objects” — with absolute precision. To maintain operational awareness in dynamic and contested environments, the Department of War identified a requirement for adaptable, dual-use technologies that enhance logistics and decision-making resilience.
