Case Study: CDL combats insurance fraud and gains real-time customer insights with Elastic

How CDL Combats Fraud and Provides Customer Insights with the Elastic Stack

CDL is a UK insurance data and comparison business (≈£53M turnover, 650k hub quotes/day, 640+ staff, 50+ insurer partners) that faced a growing industry problem: quote manipulation and insurance fraud estimated at £1.3bn detected and £2.1bn undetected. To protect customers and partners CDL needed real‑time querying at scale to catch fraud, enrich lookups and generate actionable customer insight.

CDL built "Hummingbird," a real‑time enrichment and lookup service on cloud using open software and the Elastic Stack, machine‑learning anomaly detection, Kibana dashboards and supporting tools (Java, Drools, Vault, AWS/GCP, Jenkins). In production it processes ~1.7M requests/day (1,202/min), indexes millions of quotes and MOT records, and delivers fast lookups, scalable fraud detection and richer customer analytics — improving speed, reducing risk and enabling ongoing ML‑driven detection.

CDL

Matt Houghton

Technical Consultant