MercadoLibre, Latin America's e-commerce and fintech giant serving 140 million active users across 18 countries, battled sophisticated fraud networks exploiting the platform. Fraudulent sellers, stolen credit cards, and account takeovers cost the company $340 million annually while eroding buyer trust. Traditional rule-based detection couldn't keep pace with fraudsters who adapted tactics weekly. The company needed AI that could detect emerging fraud patterns in real-time across payments, marketplace listings, and user accounts.
MercadoLibre deployed machine learning models analyzing billions of marketplace listings, payment transactions, user behaviors, and device signals to detect fraud across its ecosystem. The AI system scored every transaction, listing, and account action in real-time, blocking suspicious activity before financial loss. Graph neural networks identified fraud rings by analyzing connection patterns between accounts, devices, and addresses. Natural language processing detected fraudulent product listings and phishing attempts. The platform shared fraud intelligence across its payments arm (Mercado Pago) and marketplace operations.
“AI fraud defenses are essential infrastructure for Latin America's digital economy. We protect millions of users while enabling seamless commerce across the region.”— Chief Risk Officer, MercadoLibre
This case study is based on publicly available information about MercadoLibre.
Learn more about MercadoLibreEvery transformation starts with a conversation. Let's discuss your challenges and opportunities.
Discuss Your Challenge