PNRPU developed an AI system for bankruptcy prediction with 94.45% accuracy

As TASS reported, PNRPU has created an early warning system for companies’ financial crises based on artificial intelligence. It is designed to identify in advance enterprises that are in the risk zone.

The model uses open accounting statements and creates a three-year financial profile for each enterprise. In its analysis, it takes into account not only current indicators, but also their dynamics over time.

To train the system, the developers used reporting from Russian organizations for 2019–2024. The final sample included more than 2.3 million financial profiles of enterprises, and the development itself does not require supercomputers and can run on standard computing infrastructure.

The university believes that the solution is suitable for monitoring a large set of enterprises and identifying companies that need increased attention even before a pronounced crisis. It is proposed for use by regional ministries of economy, banks, insurance companies, and the enterprises themselves.

PNRPU also notes that the technology can be applied to monitor financial stability in industry, trade, and services. According to the developers, the system makes it possible not only to rank companies by risk level, but also to form a basis for management decisions.