Consider A TAS for testing a desktop application via its GUI. All the test cases of the automated test suite contain the same identical sequences of steps at the beginning (to create the necessary objects when doing a preliminary configuration of the test environment and at the end (to remove everything created - specifically for the test itself during the preliminary configuration of the test environment). All automated test cases use the same set of assertion functions from a shared library, for verifying the values in the GUI
fields ( e.g text boxes).

What is the BEST recommendation for improving the TAS?