Integration of artificial intelligence in banking IT infrastructure: technical and financial aspects

The use of artificial intelligence (AI) in the banking sector is becoming one of the main mainstream. Almost daily there are reports about the introduction of AI in one or another bank out of the top hundred. At the same time, medium and small banks are not yet ready for large-scale use of AI. Despite active discussions about the strategic importance of AI and machine learning, there is no research on the digital transformation of these technologies into banks' business processes. Most of the work focuses on general theoretical issues and the effect of AI implementation. The assessment of the cost of creating AI often remains hidden or underestimated. The purpose of the study is to determine the possibilities of using AI technologies in banks with a limited IT budget, to decompose the AI implementation process, identify the main stages and estimate the cost of creating a model. General scientific methods are used – analysis, synthesis, abstraction. According to the results of the study, a scheme of the process of creating and implementing AI in a bank is proposed; a characteristic of each stage is given; an estimated calculation of the cost of creating a model is made; it is proved that the proposed model for building AI in a single bank is not something supercomplicated; the costs of building an AI model are quite high, but in this case banks can quite afford to join in partnership with each other or with FinTech. The results of the study are new and of practical importance for medium and small banks when making decisions on the creation and implementation of AI models.