Case Study: CEMIG achieves US$420,000/month savings and 50%+ fraud detection with SAS Enterprise Miner (SAS)

Combating energy theft with analytics

Cemig, a major electric utilities provider in Brazil serving more than 7.5 million customers across Minas Gerais, faced widespread nontechnical and technical energy losses that posed safety risks and significant financial damage — Brazil’s regulator estimated nontechnical loss at about US$1.2 billion annually. With millions of meters to monitor, Cemig needed a way to detect theft, billing errors and meter failures more efficiently than random physical inspections allowed.

Cemig implemented SAS Enterprise Miner to build statistical models using consumption history, socio‑demographic and geographic data, generating risk scores to prioritize inspections. The analytics-driven approach raised detection rates from about 35% to more than 50%, pinpointed device failures and high‑offset consumer units, and delivered roughly US$420,000 in monthly savings.

CEMIG

Iguatinan Monteiro

Planning Engineer