TESTAR: Applying AI and Machine Learning to Testing Applications
During the three years of the project, Marviq’s developers and engineers have worked to advance the TESTAR application to be used for online testing. At its core, the tool is designed to check the source code of desktop, web, and mobile applications. In addition, it can also check whether the links contained in the application work as intended.
Our experts have programmed the tool and implemented machine learning techniques so that it can get better and more accurate over time. TESTAR simulates human behavior when checking applications very accurately, minimizing the number of random actions. By working in a way that mimics human behavior but is improved by constant self-learning, TESTAR can detect and point out most errors and mistakes that would otherwise be difficult to find.
A more advanced version of the AI would be able to check the correctness of links and other objects such as images or videos. Ultimately, it would be possible to use AI in such a way that the app deliberately tries to find the website’s weak points to break it – thus imitating the work of a professional tester.
By allowing TESTAR to learn and constantly adapt while behaving human-like, it is possible to test the source code and discover any bugs. At the current stage of development, the application cannot fully replace a human tester, but it contributes significantly to the quality and speed of online testing.