I am currently working for the Test Team of my company and one of the managers gave me the task, to reduce manual regression tests based on code coverage. Now, before anyone mentions, that manual tests should be chosen depending on Use Cases and Requirements, please consider, that this task was not mine to choose, but instead a task i have to solve. The tested application is deployed obfuscated via click once deployment and since this makes it seem alsmost impossible for any application to get coverage out of a test run, i would also love an explanation how the impact anylsis works directly. Also a short explanation on how to set up Test Impact Analysis für TFS 2017 would be really appreciated, since the Microsoft Documentation doesn't really explain TIA for manual testing properly.
Manual Desktop Application Tests is not in the supported scenarios with the TIA. See Impact Analysis supported scenarios for detials.
Refer to Speed up testing by using Test Impact Analysis (TIA) to know how the TIA works and how to set up Test Impact Analysis.
Related link : TIA overview and VSTS integration