The increasing volumes of Internet information and rapid development of social networks make the problem of automated text processing more and more topical. We have studied the use of link grammar for the Kazakh and Turkish languages and considered the possibility of creating dictionaries in these languages and connecting them to the LGP system. The most interesting stage of text analysis is semantic analysis. Its main goal is to represent the meaning of texts. In this paper, we are exploring the possibility of applying the rhetorical structure theory to the Kazakh language. Some of the formal features of rhetorical relations have been described. Statements about the properties of these features have been formulated. In the future, we are planning to build a system for identifying the topic of a text and an automatic summarization system based on the results received. We believe that even partial implementation of a semantic annotation can increase the overall performance of these systems.