Abstract: A large number of tests are performed on a service automation system. An algorithm for detecting redundant test failures is provided. Each failed automated test has a set of attributes that allow for determining a degree of similarity with previous failed tests. Each failed test generates an exception containing a message and a stack trace. A message, a stack trace, a screenshot and other attributes produced by the test at the point of failure are included into a knowledge base of failed tests. This data can be analyzed for determining a cause of the test failures. The actual cause, once determined, is saved into a database and linked to a particular test failure. In order to find the redundant test failures, the attributes of a current test failure are compared to corresponding attributes of the previous test failures. The redundant test results and their causes are discarded from further test analysis.