The problem with this is that a person who makes a mistake when doing something is much more likely than someone else to make a mistake when evaluating that same piece of work.
The really big wins happen when you get a framework in place that gives you a formal language like Gherkin for expressing what needs to be tested. That's a win because it starts giving product a way to specify requirements that are both clearer and more explicitly test-like. That starts integrating test design into the spec process as opposed to having test design happen during the implementation process.