The article is devoted to the theoretical understanding of the qualitatively new trends in the development of customs technologies related to digitalization. We are talking about the prospects of creating artificial intelligence designed to identify homogeneous groups of goods during customs control. A critical assessment of existing software products used to identify goods for customs purposes is given, a number of shortcomings are noted that do not allow for full identification with the particularities of building a single Commodity Nomenclature for Foreign Economic Activities of the Eurasian Economic Union (CN for FEA of EEU). A model for the formation of databases of identification characteristics of homogeneous groups of goods and the principle of developing scenarios for expert systems for the identification of products based on artificial intelligence are proposed.