Integrated Assessment of the Economic Efficiency of AI Agents in Sales and Marketing of Small and Medium-Sized Enterprises

The study is relevant because of the shift from using generative artificial intelligence as an auxiliary tool to deploying AI agents capable of independently performing sequences of actions in customer relationship management systems, advertising platforms, and corporate knowledge bases. For small and medium-sized enterprises, such automation creates opportunities to increase sales and productivity, while raising requirements for assessing total cost of ownership, output quality, and technological risks. The purpose of the study is to develop a method for integrated assessment of the economic efficiency of AI agents in sales and marketing processes. The research uses a mixed-method design combining content analysis of publications from 2021–2026, business-process decomposition, and scenario-based financial modelling of a representative small enterprise. A risk-adjusted total-effect model is proposed that incorporates incremental contribution margin, the value of released working time, marketing cost savings, improved customer interaction, operating expenses, and expected losses caused by agent errors. The scientific novelty lies in the introduction of a useful-autonomy coefficient reflecting the proportion of tasks completed without rework, employee acceptance of outputs, and compliance with established rules. In the baseline scenario, the model case produces a first-year net effect of RUB 2.13 million, a return on investment of 60.1%, a payback period of 7.5 months, and a three-year net present value of RUB 6.14 million. The results show that conversion growth under controlled autonomy, rather than headcount reduction, is the main driver. A phased implementation approach based on piloting, benchmark measurement, and subsequent scaling is recommended.