A Comprehensive Model for Assessing the Organizational and Economic Efficiency of Business Process Automation Using AI Agents

Abstract

This article proposes a comprehensive model for assessing the organizational and economic efficiency of business process automation using AI agents. It substantiates the inadequacy of traditional labor cost savings calculations, as agent systems simultaneously generate direct, indirect, and risk-dependent effects, as well as require costs for integration, computing resources, data, monitoring, and managing organizational changes. A multi-level system of metrics has been developed, including economic, process, quality, risk-based, and organizational frameworks. A total cost of ownership model, a procedure for calculating risk-adjusted cash flow, and an integrated performance index for agent-based automation have been developed. The performance assessment model was applied to a simulation scenario for automating the end-to-end process of processing contractual applications at a defense industry enterprise in the Ryazan Region. It has been established that the positive local effect of an individual AI agent does not guarantee an increase in the efficiency of the overall process unless the assessment takes into account the costs of verifying results, handling exceptions, resolving errors, and redistributing personnel functions.