Name in original language
Kaşif
Initiative overview
KAŞİF was developed to address the increasing scale, complexity and speed of financial and tax data that traditional rule-based audit systems could no longer effectively process. The initiative leverages big data technologies, artificial intelligence and machine learning to analyse daily data flows covering taxpayer registration, invoices, accruals, collections, customs, declarations, import–export activity and business capacity information for more than four million taxpayers.
The core objective of the initiative is to strengthen the fight against the informal economy and prevent public revenue losses by enabling early, behaviour-based risk detection. Approximately 260 calculated attributes per taxpayer are used as inputs to advanced analytical models that dynamically identify abnormal patterns, sudden behavioural changes and high-risk profiles without relying solely on predefined rules. This allows the system to detect risks such as fake invoice schemes and unjustified VAT refund claims at very early stages.
KAŞİF is fully integrated into operational processes within the Revenue Administration Directorate. Risky cases identified by the system are automatically transmitted to relevant tax offices and inspection units, enabling rapid initiation of audits, field inspections or administrative actions. The initiative operates on TÜBİTAK BİLGEM’s national Safir Cloud and Safir Big Data platforms, ensuring security, scalability and data sovereignty. In the future, the system is expected to expand with additional data sources and models, further institutionalising AI-driven decision support in tax administration.
