THE IMPACT OF ARTIFICIAL INTELLIGENCE ON CHATBOT-DRIVEN PERSONALIZATION IN MARKETING COMMUNICATIONS

Abstract

The article explores the specific features of personalized marketing communications that involve artificial intelligence and next-generation chatbots. The aim of the study is to analyze the typology of modern chatbots and evaluate the effectiveness of their application within personalized marketing strategies. Methodology and methods: the study employed general scientific methods of cognition, including analysis and synthesis; induction and deduction; comparison; generalization; a systems approach; and the modeling method. The research findings show that modern chatbots – especially those based on models like ChatGPT – are advanced technological agents capable of imitating human communication through the use of natural language processing (NLP), machine learning (ML) algorithms, and transformer architectures. Within the proposed typology, chatbots are classified as button-based, scenario-driven, AI-oriented, and messenger-integrated, each serving specific functions depending on business objectives and the selected marketing strategy. From a technical standpoint, three primary chatbot architectures are identified: generative (GPT); retrieval-based (RAC); and hybrid models that combine accurate information retrieval with the adaptability of language modeling. The study also highlights the drawbacks of AI-driven personalized communication, which include generating irrelevant responses; overlooking emotional context; and having limited ability to tailor communication styles to individual users. These shortcomings can lead to a loss of consumer trust, reputational risks, and legal implications associated with breaches of ethical standards and regulatory requirements, including provisions of the EU Artificial Intelligence Act. The research concludes that A/B testing serves as a key tool in mitigating such risks by enabling empirical comparison of the effectiveness of different personalized interaction strategies. This approach makes it possible to flexibly tailor chatbot responses to the needs of specific audience segments based on factors such as tone, formality, emotional nuance, and contextual depth. The practical contribution of this research lies in offering a structured framework and applied recommendations for enhancing the effectiveness of AI-powered personalized marketing communications.