Experian Improved Agent Response Times Under 3 Seconds with Snorkel

Experian's customer support agents were manually finding and customizing response templates for common questions, and leadership wanted faster responses so agents could focus on complex requests. It built an AWS-based agentic system that redacts PII from incoming emails, classifies them for automation-suitability, and uses a Meta Llama3-8b model fine-tuned on Snorkel-curated data (queried multiple times across functional modalities) to draft replies, which are then routed to human agents for review and sending. The result: 1-3 second agent response times, 35% of customer email responses now automated, and an 8% improvement in net promoter score.

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.