In the context of the digital transformation of sports, there is an increasing need for systematic use of data and artificial intelligence (AI) technologies to support decision-making at all levels – from individual athlete training to the management of sports infrastructure. The aim of this paper is to justify a conceptual architecture of a digital sports platform based on the integration of data analytics and AI methods and adapted to the conditions of the Republic of Kazakhstan. The study employs content analysis of scientific and applied publications, a comparative analysis of the functionality of existing foreign and Russian software solutions for sports, as well as an expert assessment of their applicability in the national context. Based on the analysis, a set of requirements for a digital sports platform is formulated, including the integration of heterogeneous data sources, support for different levels of sports training, adaptive AI-driven analytics, scalability and openness of the architecture. A conceptual architecture of the platform is proposed, comprising data sources, data storage and processing, AI-based analytical modules, user services and integration with external information systems, while ensuring information security. The implementation of such a digital platform can enhance the efficiency of the training process, support the development of mass sports and enable more evidence-based sports policy planning in Kazakhstan.