Case Study: Santéclair achieves 3x more effective fraud detection with Dataiku

How Santéclair Leverages its Data with Dataiku to Accurately Identify Fraudulent Claims

Santeclair, a subsidiary serving more than 10 million beneficiaries for supplementary health insurers, was struggling to keep up with increasingly sophisticated fraud in reimbursements from both providers and patients. With over 1.5M reimbursement claims a year and a manual audit process driven by static "if-then-else" rules, their investigation team spent too much time on low-risk cases; they turned to Dataiku (using Dataiku Data Science Studio/DSS) via IMT TeraLab and partner Eulidia to modernize fraud detection.

Eulidia used Dataiku DSS to build automated, continuously retrained machine‑learning models that combine hundreds of variables (patient/prescriber history, interaction graphs, prescription features, etc.) and operationalize predictions for audit teams while enabling Santeclair to internalize data skills. The Dataiku solution made fraud detection teams three times more effective, moved a POC from weeks to production in months with minimal IT impact, provided automated model monitoring to prevent drift, and materially reduced fraudulent payments and costs for Santeclair’s customers.

Santeclair

Jocelyn Philippe

Head of Partnerships and Development