Generative AI-Powered Self-Healing Framework for Automated Software Testing

Automated software testing plays a critical role in ensuring software quality, reliability, and rapid delivery in modern Agile and DevOps environments. However, traditional automation frameworks frequently fail due to dynamic changes in web application interfaces, resulting in broken locators, unstable test scripts, and increased maintenance effort. To address these challenges, this research proposes a Generative AI-powered self-healing framework for automated software testing using Playwright and TypeScript. The proposed framework integrates automatic locator healing, dynamic XPath and CSS regeneration, AI-based DOM analysis, failed test recovery, intelligent wait handling, screenshot and log analysis, and LLM-based test repair suggestions. The system continuously monitors test execution failures, analyzes DOM changes, and generates adaptive recovery strategies to maintain automation stability. By leveraging Generative AI and semantic element analysis, the framework reduces manual intervention and improves test execution reliability in dynamic web applications. Experimental evaluation demonstrates improved healing accuracy, reduced maintenance overhead, enhanced synchronization handling, and faster recovery from UI changes compared to traditional automation frameworks. The proposed approach contributes toward building intelligent, resilient, and scalable next-generation software testing systems.