Grades aren't necessarily an indicator on if a person comprehends the educational material. Someone can visibly under-perform on general tests, but when questioned in-person/made to do an exam still recite the educational material from the top of their head, apply it correctly and even take it in a new direction. Those are underachievers; they know what they can do, but for one reason or another, they simply refuse to show it (a pretty common cause tends to be just finding the general coursework to be demeaning or the teachers using the wrong education methods, so they don't put a lot of effort into it[0].) Give them coursework above their level, and they'll suddenly get acceptable/correct results.
IQ can be used somewhat reliably to identify if someone is an underachiever, or if they're legitimately struggling. That's what the tests are made and optimized for; they're designed to test how quickly a person can make the connection between two unrelated concepts. If they do it quick enough, they're probably underachieving compared to what they actually can do and it may be worth trying to give them more complicated material to see if they can actually handle it. (And conversely, if it turns out they're actually struggling, it may be worth dedicating more time to help them.)
That's the main use of it. Anything else you attach to IQ is a case of correlation not being causation, and anyone who thinks it's worth more than that is being silly. High/Low IQ correlates to very little besides a sort of general trend on how quickly you can recognize patterns (because of statistical anomaly rules, any score outside the 95th percentile is basically the same anyways and IQ scores are normalized every couple years; this is about as far as you can go with IQ - there's very little difference between 150/180/210 or whatever other high number you imagine).
IQ can be used somewhat reliably to identify if someone is an underachiever, or if they're legitimately struggling. That's what the tests are made and optimized for; they're designed to test how quickly a person can make the connection between two unrelated concepts. If they do it quick enough, they're probably underachieving compared to what they actually can do and it may be worth trying to give them more complicated material to see if they can actually handle it. (And conversely, if it turns out they're actually struggling, it may be worth dedicating more time to help them.)
That's the main use of it. Anything else you attach to IQ is a case of correlation not being causation, and anyone who thinks it's worth more than that is being silly. High/Low IQ correlates to very little besides a sort of general trend on how quickly you can recognize patterns (because of statistical anomaly rules, any score outside the 95th percentile is basically the same anyways and IQ scores are normalized every couple years; this is about as far as you can go with IQ - there's very little difference between 150/180/210 or whatever other high number you imagine).