My first knee-jerk reaction to your post was the same as the people who downvoted you: Here's another person incorrectly implementing a process, and then blaming the process when it fails.
But, real talk: Automated testing is borderline impossible without buy-in from the developers maintaining the code under test.
You get developers changing function signatures so tests no longer compile. You get changes to function invariants that invalidate the tests. You get UI changes that break tests that look for component placement, or that search for certain strings in the UI.
If your developers insist on changing these things (or are forced to change these things by constantly-shifting project requirements) without updating the tests, then you are just treading water.
In that case, you probably are better off with manual tests.
But, real talk: Automated testing is borderline impossible without buy-in from the developers maintaining the code under test.
You get developers changing function signatures so tests no longer compile. You get changes to function invariants that invalidate the tests. You get UI changes that break tests that look for component placement, or that search for certain strings in the UI.
If your developers insist on changing these things (or are forced to change these things by constantly-shifting project requirements) without updating the tests, then you are just treading water.
In that case, you probably are better off with manual tests.