Alice — Brazil's AI system for detecting procurement fraud
Brazil
Alice, deployed by Brazil's CGU since 2015, uses NLP and RPA to screen federal procurement notices daily, reducing audit turnaround …
Romania · Bucharest · See the Romania profile
Romania's ANI automatically cross-references procurement officials' identities with registered bidders before contracts are signed; over 100,000 procedures worth RON 72 billion screened, €108 million in suspect contracts blocked, undisclosed conflicts fell 52% within 2.5 years.
PREVENT is an automated ex-ante system, launched in June 2017 and operated by Romania's National Integrity Agency (ANI), that cross-references the identities of officials involved in public-procurement procedures on the SEAP platform against National Trade Registry and Personal Records data to detect undisclosed conflicts of interest with bidders before contracts are signed.
Since launch the system has screened over 100,000 procurement procedures worth approximately RON 72 billion (EUR 15.5 billion), issued 57 integrity warnings, and prevented 55 contracts worth EUR 108 million from being awarded in potentially conflicted conditions; the rate of undisclosed conflicts fell by 52% in the system's first 2.5 years of operation. A second-generation PREVENT 2.0, operational since February 2026, screened 6,414 procedures worth 86 billion lei in its first quarter, identifying integrity risks totalling 2.2 billion lei.
The European Commission's Anti-Fraud Knowledge Centre features PREVENT as a good practice, and the OECD's 2025 AI-in-public-procurement report cites it as a model for AI-assisted integrity in contracting.
Implementation detail (cost, timeline, staffing, conditions for success) is not yet available for this practice.
Where this practice's information was retrieved from, and when.
Brazil
Alice, deployed by Brazil's CGU since 2015, uses NLP and RPA to screen federal procurement notices daily, reducing audit turnaround …
Malaysia
Bank Negara Malaysia (BNM) and PayNet's National Fraud Portal (April 2024) links 53 institutions via FNA graph analytics to trace …
Ireland
Since 2011, Ireland's Revenue Commissioners have used AI risk scoring (REAP) to target tax-compliance interventions; 290,000 data-analytics-driven interventions in 2023 …
United States of America
The US Treasury's Office of Payment Integrity uses ML to screen federal disbursements. In FY2024, ML-driven processes prevented or recovered …
Open full copilot Grounded in cited practices — always check the sources.