Resume. Introduction. The aging of the population, accompanied by an increase in age-related diseases and cognitive impairment, is driving the search for innovative solutions in providing care for the elderly. Traditional approaches to creating a safe, barrier-free therapeutic environment, encompassing a range of preventive, therapeutic, and rehabilitative measures, require further development, including the use of artificial intelligence. Artificial intelligence technologies, including voice assistants and monitoring systems, offer the potential for personalized care, particularly for individuals with cognitive, sensory, and motor impairments. However, the use of these technologies in long-term care facilities remains understudied, making this study relevant. Objective. To evaluate the potential of using artificial intelligence (AI) using a voice-activated AI agent in a therapeutic environment for elderly individuals with pre-dementia and mild cognitive impairments. Materials and methods. The study included 44 patients (16 men, 28 women) aged 65 to 75 years in a long-term care facility. The average age was 71.6 ± 5.5 years. Inclusion criteria: age 65 years and older, MMSE 20-25 points, intact hearing sufficient for speech perception, absence of severe speech impairments, voluntary informed consent. To create an adaptive environment for patients in the main group, which included 24 patients, a voice AI assistant was used, providing, in accordance with the doctor's recommendations, reminders for taking medications, informational and emotional support, and cognitive stimulation. At the initial stage and after 12 weeks of the study, cognitive abilities were assessed using the MMSE scale, psychological state using the PSM-25 scale, and a survey was conducted using the developed questionnaire. Results and discussion. The study revealed a significant reduction in psychological stress indicators in patients in the study group interacting with the voice AI agent. Low stress levels were reported by 75% of the subjects, and 66.7% of the elderly reported an improved mood (p