Those rejection emails are most likely sent by an AI. If you reply back and ask them to specify exactly what is wrong, you'll get the same generic email back. Ask again, and they'll send the same generic response without any details or comments written by a human. They simply can't specify the problem at all. That's how you know you are talking with an AI, not a human.

The correct way to respond to those rejection emails is to ask for a "human being" (this is the keyword that works) to review the case. Also explain in the email why there isn't anything more you can do (if you have done every possible fix already).

As a side note, when AI systems get more common, this will be a common nightmare for regular people. When an AI makes an incorrect decision regarding you, no-one can check the code why it happened because the code doesn't exist. All we may have are some weighted matrices and neural network data as bunch of numbers.