If you discovered that an automated test case was consistently failing due to an issue in the application, what steps would you take to resolve it?