Abstract
The digitalisation of the banking sector is changing the way borrowers are assessed, loans are granted, and debt is being tracked. At the same time, the expansion of remote services is accompanied by the risks of accelerated accumulation of debt burden, misclassification of clients and fraudulent loan processing. The purpose of the article is to assess how digital technologies affect banks' debt problems in Kazakhstan and to identify indicators to verify the effectiveness of their use. The study is based on official data from the Agency of the Republic of Kazakhstan for Regulation and Development of the Financial Market for 2021-2025, payment statistics from the National Bank, an analysis of regulatory measures, and international scientific literature. The study applies dynamic and structural analysis, calculates the rate of debt change, and compares technological tools by stages of the credit process. It was found that the volume of loans overdue for more than 90 days increased from 669 billion tenge at the end of 2021 to 1,575 billion tenge at the end of 2025. In 2025, its growth reached 44.0%, while the loan portfolio grew by 21.3%; the share of NPL90+ increased from 3.1% to 3.6%. The paper shows differences between household and business debt, as well as between payment digitalisation and automation of credit risk management. A scheme for the interaction of scoring, early warning, biometric identification and digital debt settlement has been developed. The scientific contribution consists in combining stock analysis and relative debt indicators with a system for verifying technological results. Suggestions for assessing the quality of new outputs, sustainability of restructuring, and independent validation of models are of practical importance. The results obtained characterise sectoral trends; the causal effect of individual technologies requires data from banks and loan agreements.