As part of the analysis of management procedures for multi-agent systems, the need to introduce the effect of self-organization into them is shown and the nature of the emergence of swarm intelligence in various spheres of social and industrial life is investigated. In particular, it is in sociology, economics, transport, medicine, and other sectors of the national economy. It is shown that the basis of self-organization and, as a result, the emergence of collective intelligence, is a multi-agent interaction of differently structured subjects. A finite state machine is chosen as the main model of the subject’s behavior (interaction agent). The adaptation of agents’ collective to the changing environmental conditions is carried out through the generation and selection of new generations of state machines. It is shown that the adopted model finds its confirmation in real nature (using the examples of colonies of ants, termites, bats, etc.). The need to create a digital platform that integrates all the necessary information about the system, which forms possible options for the development of a multi-agent system based on simulation modeling, which makes it possible to search for suboptimal solutions, is substantiated. The created mathematical model for describing this natural, collective intelligence can be effectively used in creating various artificial intelligence systems (for example, in managing the socio-economic development of a region based on a single information space; in the carriage fleet management system of the country or region).