In theory, sure. In practice, to do this you'd need to find a mostly foolproof way of knowing who was "good" when they signed up*. Will the team reviewing be able to KNOW who is likely to be good? if not, you're just providing friction for signups.

This is a perfect thing to test. I'd use cohort analysis to follow a group you let in vs the control and see if the team does a better job than average at filtering out the "bad."

The Test would also then tell you if that friction point was hurting you overall, even if the team was unable to manually discern who the best customers are.

Another option: just review all the signups manually anyway and disable the ones you don't like. No need to reveal it.

*To do this, you'll need to find out what your good customers have in common. (the same industry? title? company growth phase?). This is called segmentation. Once you understand that you can then market to get to more of the good.