AI Helps Detect Violations in Public Procurement
Russia’s Federal Antimonopoly Service (FAS) has launched its first administrative proceeding after an AI system independently detected signs of a cartel agreement.
The Russian State Information System Antikartel (Antikartel, Anti-Cartel System), designed to prevent and detect violations of competition law, added artificial intelligence capabilities in 2026. The system runs on the GosTekh (State Technology) platform and uses risk-based analytical methods, machine learning, and a large language model. FAS operates the platform, which consolidates procurement and bidding data from government information resources. Antikartel analyzes large datasets to identify suspicious patterns. Previously, analysts carried out this work manually. Now, with AI integrated into Russia’s government digital infrastructure, algorithms handle the analysis.
The neural network detected signs of possible violations in the actions of two companies, Stroyprofigrupp LLC and MT Grupp LLC. FAS suspects that the companies broke the law during tenders for building maintenance at children’s educational institutions in the Odintsovo urban district of the Moscow region. According to the agency, the companies gave the appearance of competition to keep prices where they wanted them. If the violations are confirmed, the parties involved in the alleged cartel could face administrative fines.
Further Integration and Adaptation
FAS plans to continue its digital transformation program and introduce new intelligent algorithms. In the future, Antikartel could be integrated with other state information systems, allowing the agency to obtain the information it needs without submitting additional requests. Procurement analysis automation is also expected to expand into commodity markets, particularly the pricing of goods considered socially important. The Russian government announced these plans when the system was first developed.
Antikartel was designed specifically for Russia’s domestic infrastructure, so selling the platform in foreign markets is not currently under consideration. However, some of its components could attract export interest. These might include algorithms for analyzing procurement data and detecting anomalies, methods for assessing the risk of collusion, software components that integrate information systems, and expert services for developing and deploying digital antitrust oversight tools. Russian products would still have to compete with systems already on the market. To enter foreign markets, domestic solutions would need to be adapted to each country’s laws, procurement data structures, and investigative procedures.
Russia’s System Reflects a Global Trend
A law establishing the Antikartel state information system was signed on June 24, 2025. It set the legal framework for the system and designated Russia’s FAS as its operator. By December of that year, the government had reported the first results: the system had automatically analyzed 9.6 million tenders against 20 technical and behavioral criteria. In 2025, FAS specialists used Antikartel to uncover 21% of cartels in public procurement. By May 2026, that share had risen to 40%.
Other countries are also automating public procurement analysis. Denmark, for example, uses Bid Viewer, a system that applies statistical indicators, bidder behavior analysis, and machine learning models to identify suspicious tenders. In Brazil, the antitrust agency CADE developed Cérebro, which combines data from public and private sources and analyzes similarities in bidder behavior. It also uses statistical methods to detect attempts to simulate competition.
In September 2024, the Organisation for Economic Co-operation and Development (OECD) published a report examining methods that allow antitrust authorities to identify potential violations on their own, including through new digital tools. The report’s authors concluded that effective cartel detection depends on combining different methods.
Russia’s project aligns with the international shift toward digital oversight, and intelligent algorithms are likely to play an increasingly important role in detecting violations.
Stronger Competition, Fewer Budget Losses
AI is increasingly becoming a practical tool for government oversight in Russia. In FAS investigations, additional data sources could be brought in, while more sophisticated machine learning models may help identify not only known forms of suspicious behavior but also new ones. Still, false positives remain possible: unusual prices or repeated wins by the same companies may reflect the specific conditions of a market. An automated alert, therefore, should not be treated as proof of wrongdoing.
Even so, the tool could help foster fair competition, particularly in capital-intensive industries such as construction. The system’s performance will ultimately be measured by the number of confirmed violations, the budget losses it prevents, and its ability to protect the rights of all bidders.
