INTELLIGENT MANAGEMENT SYSTEM FOR DIGITAL MARKETING COMMUNICATIONS BASED ON ARTIFICIAL INTELLIGENCE AND CHATBOTS
Topicality. The article substantiates a conceptual approach to optimizing and ensuring cybersecurity of digital communications management systems of enterprises by forming a single integrated information space of a business entity. The relevance of the study is due to the rapid deployment of artificial intelligence (AI), which has led to the emergence of specific cyber threats and risks of compromising confidential information at the points of infrastructure integration of linear MarTech platforms. The subject of the research is methods, models and tools for building an intelligent digital marketing communications management system based on artificial intelligence technologies, as well as mechanisms for ensuring its cybersecurity, verifying protection circuits and assessing the effectiveness of functioning in the context of the integration of digital MarTech platforms. The purpose of the article is to develop a conceptual model of an intelligent digital communications management system with verification of its security contours and the formation of an applied mathematical apparatus for assessing its effectiveness. The following results were obtained. A four-circuit architecture of the digital communications management system has been formed, where the basic digital infrastructure circuit acts as a source of primary data for the analytical AI core, and the security of access to CRM systems is regulated by a separate feedback loop. The information flows of the system have been formalized based on a set-theoretic approach using input parameters (set X) and output control signals (set Y), which record text and voice information protocols of dialogues. A model of integral evaluation of the system has been proposed, which consolidates operational-information, security and performance parameters. The results of calculating the eigenvector of the matrix using the T. Saati MAI (under the condition of full consistency, where VU = 0.69% < 10%) allowed us to determine the objective weighting factors of the system. It was found that the cybersecurity coefficient (γ = 0.297) and the level of automation of circuits (α = 0.163), which function in synergy with the productivity index (β = 0.540), are of great importance for ensuring the reliability of the architecture. Conclusion. The significance of the work lies in determining the strategic prospects for data protection through the deployment of a zero-trust architecture and AI governance standards, which protect personal data from unauthorized fine-tuning of public neural networks during operation.