AI can now analyze real user behavior from application logs to automatically design smarter and more realistic test cases.
QA engineers are evolving into explorers who experiment with new AI tools and builders who create custom automation solutions.
Traditional testing methods struggle to simulate the unexpected and irrational ways real users interact with an application.
A four-step process transforms raw application logs into structured, production-aligned test cases using large language models.
Comprehensive user interaction data is gathered from the frontend, API gateway, and backend services to create a complete system view.
Unstructured and noisy production logs are normalized and correlated to create a coherent narrative that an LLM can interpret.
Well-defined prompts instruct the LLM to generate test cases in specific formats like JSON or Gherkin for easy integration.
A second LLM, acting as a judge, evaluates the quality and relevance of generated tests by providing scores and feedback.
Validated test cases become reusable assets that can be imported into automation frameworks and test management platforms.
The role of a QA engineer is shifting from manually creating tests to guiding and validating AI-driven test generation systems.
Jobs that call for the skills explored in this talk.