Evaluating Chatbot usability: Adaptation of the BUS-11 Questionnaire for a Russian-speaking sample

Context and Relevance. Chatbots powered by generative artificial intelligence (AI) are becoming a crucial tool in education, healthcare, and service industries. A key factor in their effectiveness is usability; however, specialized assessment tools for Russian-speaking users have been lacking until now. Objective: to adapt and validate a Russian-language version of the short questionnaire BUS for evaluating the usability of generative AI-based chatbots. Hypothesis. The Russian-language version of the tool demonstrates reliability and validity, and replicates the key factors of the original methodology: Perceived Dialogue and Information Quality, Perceived Bot Function Quality, and Perceived Function Accessibility. Methods and Materials. The study involved 207 participants (aged 18—60); test-retest reliability was assessed on a sub-sample of 98 respondents with an interval of 1—2 months. Procedures included linguistic adaptation, exploratory and confirmatory factor analysis, Cronbach's alpha coefficient, and correlation analysis. Results. The questionnaire demonstrated high internal consistency (α = .81) and test-retest reliability (r = .83). The three-factor structure was confirmed: Dialogue Quality, Functionality, and Function Accessibility. Convergent validity was established through high correlations with UMUX-LITE (r = .845) and AttrakDiff scales; divergent validity was confirmed by a low correlation with the Satisfaction with Life Scale (r = .12). Conclusions. The adapted version of the short BUS is a reliable tool for assessing the usability of chatbots on a Russian-speaking sample and can be used in both scientific and applied research.