Do you have a background in CS or optimization? MILP is a pretty standard concept in algorithms/optimization. So this example doesn't really convince me that the AI reached some unusually superior conclusion. It sounds more like it suggested a well-known technique that you personally hadn't encountered. Useful, yes, but that seems more about background knowledge gaps than about the merits of letting the tool run unconstrained.
There are always concepts that some people think are a basic, that others haven't heard of. The entire benefit here is that AI can point out what we miss. There are certainly techniques you don't know about, or just didn't think to apply to a problem, that others would find to be pretty standard.