Turkey Deploys AI System “KASIF” to Crack Down on Shadow Economy

Turkey’s Ministry of Treasury and Finance has launched a new artificial intelligence–powered system to intensify its fight against the shadow economy and fake invoicing schemes. The application, named KASIF, has been developed jointly by the Revenue Administration (GİB) Technology Department and the TÜBİTAK Artificial Intelligence Institute, marking a major technological leap in tax enforcement.

Powered by machine learning and big data analytics, KASIF is now actively scanning millions of taxpayers every day, instantly flagging suspicious transactions and reporting high-risk cases to tax authorities. Officials describe the system as a game-changer in identifying fraudulent activity before large-scale losses occur.

Rule-Free Analysis Focused on Behavioral Shifts

Unlike traditional audit systems that rely on predefined and rigid rules, KASIF operates without fixed thresholds or static criteria. Instead, it analyzes tens of terabytes of data across integrated big data platforms, evaluating taxpayer information daily.

The system processes a wide range of variables, including tax declarations, issued invoices, collections, import and export activity, employee numbers, and operational capacity. What sets KASIF apart is its ability to autonomously detect abnormal changes in commercial behavior, without requiring prior assumptions or manual input.

By comparing a company’s declared capacity with its invoicing volume, KASIF can immediately spot inconsistencies—such as businesses with no production capability issuing massive invoices or newly established firms engaging in unusually high-value transactions. These behavioral anomalies are identified in real time, significantly reducing response times.

Billions in Potential Fraud Blocked in Days

Shortly after being activated, KASIF identified numerous high-risk taxpayers involved in fake documentation schemes, preventing potential losses worth billions of lira. Authorities shared several striking examples detected by the system:

A wholesaler was flagged for issuing three invoices on the same day it registered as a taxpayer, even though it had made no product purchases.
A company was identified for issuing five fake invoices totaling ₺50 million within just five days of its establishment, targeting multiple high-risk entities.
Seven companies were found to be registered under the name of a single salaried employee across different tax offices; six of them were confirmed as shell companies, collectively issuing ₺246 million in fraudulent invoices.
In a large-scale data sweep, 2,550 taxpayers established within the last three months were flagged after issuing ₺276 billion worth of invoices despite showing no real economic activity.
A taxpayer who requested a ₺7.7 million VAT refund was identified by KASIF as making an unjustified claim, prompting immediate intervention.

These findings underscore the system’s ability to uncover complex fraud networks that might otherwise go undetected for months or even years.

How the Enforcement Process Works

Once KASIF identifies a taxpayer as high-risk, two parallel enforcement mechanisms are triggered.

First, detailed audit files are transferred to the Tax Inspection Board for comprehensive tax examinations. These reviews focus on verifying documentation, transaction authenticity, and compliance with tax regulations.

Second, the relevant taxpayers are reported to their local tax offices for on-site inspections. Field teams conduct physical verifications of business premises, and if it is conclusively determined that a company was established solely for issuing fake invoices, its tax registration is permanently canceled by commission decision.

This dual-track approach ensures both rapid intervention and thorough legal scrutiny.

Minister Şimşek: “We Detect and Act Instantly”

Treasury and Finance Minister Mehmet Şimşek said the introduction of KASIF marks a new phase in Turkey’s fight against informality. Emphasizing the role of advanced technologies, he stated:

“With artificial intelligence and machine learning, we are now detecting all unregistered activities instantly and taking action without delay. Through these applications, we will continue to combat the shadow economy effectively and prevent unfair competition caused by non-compliance. I invite all taxpayers to remain within the formal system.”

A New Layer on Existing Risk Systems

KASIF strengthens GİB’s technological infrastructure by adding an advanced AI-driven layer on top of existing rule-based systems, such as the Fake Document Risk Analysis Program (SARP) and the Risky Refund Monitoring and Analysis Program (RİTAP).

While those systems rely on predefined criteria, KASIF continuously learns from new data, adapts to emerging fraud patterns, and improves detection accuracy over time. Officials believe this hybrid model significantly enhances Turkey’s capacity to protect public revenues and ensure fair competition.