Paper presents a comprehensive analytical review and synthesis of modern approaches to the development of decision support systems (DSS) based on multi-agent system (MAS) technologies. The evolution of corporate DSS from classical centralized architectures to distributed intelligent systems is examined. It is shown that the modern corporate environment is shaped at the intersection of classical decision support systems, business intelligence, enterprise analytics, heterogeneous data integration, multi-agent systems, explainable artificial intelligence, neuro-symbolic artificial intelligence, knowledge graphs, retrieval-augmented generation, and digital twins of business processes. The evolution of decision support systems from classical model-oriented and data-oriented systems to intelligent distributed platforms based on multi-agent technologies is considered. Corporate analytics systems are analyzed, in particular ontological. A classification of strategies for integrating heterogeneous data in multi-agent DSS is proposed. It is substantiated that the increasing complexity of managerial decision-making is caused not only by the growth in information volume, but also by its structural, semantic, and organizational heterogeneity, as well as by the integration of ERP and CRM systems, data warehouses, document archives, corporate communications, and external digital environments. Under such conditions, classical centralized decision support systems reveal limitations in terms of adaptability, interoperability, local autonomy, coordination across functional domains, the handling of semi-structured and unstructured data, and the explainability of recommendations. The results of the study create a theoretical foundation for the development of adaptive distributed decision support platforms in which decisions are formed not by a single centralized module, but through the interaction of specialized agents, corporate rules, knowledge layers, human oversight, and simulation mechanisms.