In this talk, we will discuss the challenges of detecting and resolving performance issues, such as video stuttering, lag, and app crashes, and how Orange Bees has developed an AI-powered QA testing system to overcome these challenges. Combining Python, OpenCV, Object Detection, and OCR, the system validates video continuity and completeness, detects gaps, and identifies device-specific issues. With out-of-the-box report generation, businesses gain valuable insights into gaps and improve the app's user experience.

To validate the solution, we put it through rigorous testing using our own bench test. We recorded a clock on a desk for several hours using various phone devices and ran their analytics on top of that. The system was able to detect frozen frames and ensure that the recordings were complete and continuous, even when recording for extended periods. We also introduced artificially corrupted videos, introducing the exact issues that they were trying to detect, and the system was able to detect these issues perfectly.