This article examines an approach to designing the architecture of management decision support systems for entrepreneurial structures using artificial intelligence (AI) technologies. It is demonstrated that traditional decision-making methods fail to provide sufficient speed and justification for management actions in the face of high market volatility and information overload. A multi-level architecture is proposed, based on the principles of modularity, human-AI hybrid interaction, explainability, and incremental development. The system's structure, comprising four layers, is presented. Particular attention is paid to the role of the human-in-the-loop and the threshold control of decisions. Using small and medium-sized enterprises as an example, it is shown that the proposed approach enhances forecast accuracy, accelerates the management cycle, and fosters a culture of trust in analytics. The architecture ensures adaptability, transparency, and regulatory compliance, making it an effective tool for implementing AI in entrepreneurial structures.

Keywords: management decision-making, architecture, artificial intelligence, human-in-the-loop, digital transformation, intelligent support