Artificial intelligence becoming widespread at biggest Russian banks - CBR
MOSCOW. Nov 20 (Interfax) - Artificial intelligence is becoming widespread at systemically important Russian banks, possibly because they have the money to develop it at present, the Central Bank said in a consultation paper.
The systemically important banks list includes 12 lenders: UniCredit Bank, Gazprombank , Sovcombank , VTB , Alfa-Bank, Sberbank , Credit Bank of Moscow , T-Bank , Promsvyazbank , Raiffeisenbank, Russian Agricultural Bank and DOM.RF Bank.
AI is used quite heavily in most core business processes and specializations at systemically important banks, from credit scoring and personalized banking to fraud detection and customer transaction analysis. For example, 11 of the 12 backbone banks use traditional AI in credit scoring and risk management, while nine use it for customer profiling, personalization of banking product and service offerings, and other functions. They also make much use of deep learning models and generative AI, the Central Bank said.
Systemically important banks have products and services that use AI with a high degree of autonomy or without human intervention, while most respondents - over 90% of those surveyed by the Bank of Russia - said such business processes were lacking. AI, including generative models, is most autonomous in business support: consulting clients (chatbot automation), generating documents at the client's request (statements and other reports), processing and verifying images and documents from clients, personalizing cashback programs, analyzing and forecasting account balances, classifying call center requests, and routing inquiries.
In lending, AI is used for the hyper-personalization of loan offers, assessing the likelihood of default, client income forecasting, cross-selling loan products, and credit scoring. Systemically important banks said the degree of AI autonomy in scoring individuals and SMEs was approaching 100%.
Loan issuance and decision reviews are fully automated in such cases, and human intervention is limited mostly to the development of AI models, their validation, and decisions to launch them. However, human intervention remains mandatory in decision-making on complex loan products and large requests.
Insurance companies and systemically important banks are the financial market AI penetration leaders. The Central Bank survey found 60% of insurers used AI on a regular basis. In the banking sector, 24% of respondents are actively using AI, and another 19% are piloting projects. Adequate AI penetration is being seen in the payment services market (33%) and the securities market (29%).
Sberbank will more than double investments in AI development to 350 billion rubles in 2026, CEO Herman Gref said on November 19. He said AI would generate 1.4 trillion rubles of income for Sberbank in 2024-2026.
AI has so far saved VTB Group 15 billion rubles, and the figure could rise to 50 billion rubles in the next two years, VTB chief Andrei Kostin said in November. T-Bank expects a direct economic effect of "tens of billions" of rubles from AI use in 2025.
