The limits of AI agent autonomy in a bank: the economic paradox of cost reduction and loss growth
Year: 2026
Type: article
Abstract: The objective of this study is to justify the maximum level of AI agent involvement in banking operations, at which further reduction of the employee's role ceases to improve the economic evaluation of the agent's use. The methodological basis of the study includes mathematical modeling of the relationship between the degree of AI agent involvement, positive results, and the costs incurred when a repeatable error occurs. Results. It is shown that expanding the role of the AI agent reduces ongoing costs and accelerates the execution of standard banking operations, but simultaneously increases the cost of erroneous logic if it is reproduced in the flow of homogeneous operations until it is detected. It is substantiated that the benefit of delegating decisions to the agent increases with diminishing returns, while costs increase as one approaches full autonomy. Conclusions. Completely delegating operations to an AI agent should not be considered a bank's preferred option. The most economically feasible level of agent involvement is one at which the increased revenue still outweighs the increased costs associated with repeatable errors.
Source: Sibirskaya finansovaya shkola
Author: T. Zverkova
Institution: Orenburg State University
Language: English
Domain: Social Sciences
Open Access status: hybrid
OpenAlex ID: W7201860786