How SLB uses Snorkel Flow to enhance proactive well management

SLB, operating in 120+ countries, needed to extract 18 industry-specific entities from unstructured drilling and maintenance reports, but varied document formats, poor domain-expert/data-scientist collaboration, and 1-3 hours of manual labeling per document were bottlenecks. Using Snorkel Flow, the joint team built a classification task hitting 85% F1 within three days, then improved it to 91.4% F1 through programmatic labeling and model-guided error analysis. Report processing time dropped from 1-3 hours to seconds (a 99% reduction), generalization improved 47%, and the model handled 15 different document structures; Swaroop Kalasapur, Head of SLB's Technology Innovation Center, said "What would have taken us months to go through an iteration can happen in minutes now. Literally!"

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.