It is common practice to write many unit tests and API tests and only write a few User Interface (UI) tests. Why? Because UI tests are brittle. If you change one thing, the other tests unravel. What if you could use machine learning (ML) to help? Many apps have the same functionality, such as login, checkout, share, and pay. In this session, Godfrey Nolan talks about how to use Object Detection and labeling techniques to make UI tests more robust with a fraction of the code.