Brazil's AI system “Alice” audits public contracts, delivering R$1.25 bn in savings

In 2024 Brazil’s federal accountability office (CGU) deployed an artificial‑intelligence tool called Alice (Analyser of Licitations, Contracts and Notices) to scan public‑procurement data. The system reviewed more than R$30 billion in purchases, flagged risks and triggered 212 preventive audits, resulting in R$1.25 billion of financial benefits for the government.

Alice operates on a daily automated cycle: it gathers contract data from multiple sources, applies text‑mining and machine‑learning techniques to detect anomalies such as overpricing, supplier favoritism or fraud, and immediately alerts auditors so they can intervene before a contract is signed. Software architect Douglas Leonardo Nunes Santana, who led the project's technical design, said, "Em auditoria pública, não basta que a IA aponte um risco; é necessário demonstrar de onde veio a informação, quais documentos foram analisados e como o auditor poderá revisar aquela indicação." He added that AI should support, not replace, human decision‑making and that strong governance—privacy safeguards, traceability and legal compliance—is essential.

The Brazilian initiative is part of a broader open‑contracting movement, echoing systems such as Ukraine’s ProZorro and Chile’s ChileCompra, which aim to increase transparency and generate public‑sector savings.