Abstract
<jats:p>Background. The article examines the problem of insufficient functionality of existing low-code platforms in terms of decision support for freight transportation management. It is shown that the growth in freight traffic volume in the Russian Federation (by 5.5% in 2024), coupled with an increase in the share of short-haul routes and an acute personnel shortage (a deficit of about 1 million specialists), creates a need for new approaches to the automation of dispatching management. It has been established that modern low-code platforms are primarily focused on accounting and registration functions and do not integrate optimization algorithms with visualization tools, while Kanban boards, used in isolation, do not provide a closed-loop learning system based on expert decisions. Purpose. Development and research of the architecture of a low-code platform with hybrid intelligence that integrates Kanban boards for visualizing transportation statuses and supporting dispatcher decision-making. Materials and methods. The study employed a systems analysis approach to identify functional gaps in existing solutions. The architecture was designed using a modular approach and the BPMN 2.0 business process modeling notation. A case study method involving a regional transport company was used to verify the proposed solutions. Results. This article proposes a three-layer architecture that includes a low-code configuration layer (a Kanban board with WIP limits, a visual route editor), a hybrid intelligence layer (an ML optimization module, an LLM explanation generation module, a decision orchestrator, a simulation modeling module), and a data and integration layer.</jats:p>