If you encountered a situation where your automated tests failed due to an unexpected change in the application, how would you address it?