Proceedings of the International scientific and practical conference ―Science in the Era of Globalization‖ (May 22-24, 2026) / Publisher website: www.naukainfo.com. - Zurich, Switzerland, 2026. - 353 p.

114 UDC 004.415.5:004.896 Kutaiev Serhii Valentynovych senior lecturer Lviv State University of Internal Affairs Lviv, Ukraine TRANSITIONING TO MODEL-BASED TESTING IN DESKTOP AUTOMATION THROUGH GENERATIVE AI CAPABILITIES Abstract. Traditional approaches to desktop test automation heavily rely on manually scripting end-to-end scenarios, which results in significant maintenance overhead as the software scales. Model-Based Testing (MBT) offers a solution by defining application behaviors via abstract models, yet creating and maintaining these models remains technically challenging for complex desktop interfaces. This paper explores a methodology for transitioning to MBT in desktop automation by leveraging Generative Artificial Intelligence (AI). The proposed framework utilizes generative models to analyze desktop UI trees (via Appium PageSource) and automatically construct behavioral state machines. By dynamically generating C# execution paths and test cases from these AI-generated models, the methodology minimizes human intervention, eliminates script redundancy, and ensures comprehensive test coverage of complex, multi-platform desktop user interfaces. Keywords: model-Based Testing, Generative AI, desktop automation, Appium, state machines, C#, automated test generation, test coverage. The scalable automation of enterprise desktop solutions requires a fundamental shift from static, script-heavy testing to dynamic, model-driven architectures. In traditional automation paradigms using Appium [1] or WinAppDriver, engineers manually code specific sequences of interactions (e.g., login, navigate, fill form, save). However, when a desktop application features hundreds of interrelated screens,