An artificial-intelligence and advanced analytics initiative used by Türkiye's Ministry of Treasury and Finance has flagged more than TRY 2 billion in public spending as carrying a risk of inefficiency. The finding is not a court ruling or a confirmed determination of misconduct. The system identifies unusual patterns in large volumes of financial data and helps specialists decide which records may require closer examination.
Financial records are screened with AI
The work is based on the AI-Supported Accounting and Advanced Analytics Project developed by the Ministry of Treasury and Finance with TÜBİTAK BİLGEM. According to the ministry's project description, the initiative is designed to make more effective use of large public-finance datasets and support decision-making with analytical methods. TÜBİTAK BİLGEM says the project applies artificial-intelligence algorithms to transactions produced by public administrations and spending units.
Current reports published by Anadolu Agency and Bloomberg HT on August 9, 2026 say the system flagged more than TRY 2 billion in spending for “inefficiency risk.” That wording does not establish that every flagged payment was incorrect, unlawful or recoverable. A risk signal is better understood as a prioritization tool that directs qualified reviewers toward transactions needing additional checks.
Why the scale of BKMYBS matters
Figures reported with the announcement show that the Integrated Public Financial Management Information System, known by its Turkish abbreviation BKMYBS, serves 497 public administrations, about 95,000 spending units and 4,000 accounting units. Its user base is reported at roughly 400,000. Annual activity includes about 150 million accounting records, 18 million payment-order documents, 10 million electronic invoices and 5.5 million electronically signed documents. The system also handles around 3 million payroll calculations.
That scale creates a dataset that would be difficult to review record by record using traditional methods alone. In this setting, artificial intelligence is not the final decision-maker. It functions as an analytical layer that highlights unusual movements, recurring patterns or groups of records that may deserve closer attention. Final assessments and any administrative steps remain dependent on review by authorized specialists.
How the figures should be read
The disclosed numbers describe different aspects of the system. Transaction counts show its annual operating scale, paper and carbon figures describe digitization effects during 2025 and 2026, and the amount above TRY 2 billion refers specifically to the risk-analysis result. They should not be treated as interchangeable measures or combined into a single estimate.
Digitization and resource savings
The latest disclosure also linked the broader digital process to savings beyond risk screening. Officials said electronic workflows in 2025 and 2026 saved approximately 200 million sheets of paper and contributed to avoiding about 5,000 tonnes of carbon emissions. These figures indicate that the initiative is connected both to financial analysis and to the replacement of paper-based documentation.
The project was also announced as the winner of the artificial-intelligence category at the 11th Productivity Project Awards. Treasury and Finance Minister Mehmet Şimşek said AI and advanced analytics applications would be expanded across public financial management. The next points to watch are how flagged risks feed into formal review processes and which measures will be used to report the system's results. Until those reviews are completed, the TRY 2 billion figure should remain described as spending carrying an identified risk, not as proven waste.
