Short answer, no.
First, g-code that uses cutter compensation is different than g-code that doesn't. There are additional g-code commands (G41/G42) and the tool path is different (it is the edge of the cut rather than the tool center), which I guess you already know if you were doing it by hand. Thus, cutter compensation strategies have to start in Fusion (by enabling compensation). Secondly, while some operations in Fusion use cutter compensation, some operations do not, and some portions of an operation might use cutter compensation, and other portions (of the same operation) might not. For example, the roughing part of the operation might not use cutter compensation, while the finish pass of the same operation does. This is fine when the difference between expected and actual tool diameter is a thou or two, but not if we are talking the difference between a 3/8" and 1/2" end mill. In the roughing operation, fusion will not be using a compensation strategy nor issue the compensation g-code commands, and will be driving a 1/2" end mill down a path where it expected a 3/8" end mill. Obviously, it will cut way too much (if not break).
The point is, even though the strategy you mention might work with something simple and hand coded, it will not work when fusion is doing the g-coding because it uses the compensation strategy only when enabled, only when available, only when needed, and based on the idea that we are talking a thou or two, not a 3/8" end mill versus a 1/2" end mill. Even though what you say is technically possible, Fusion simply doesn't think of "cutter compensation" in the same sense. It thinks of it as something to use at the end of the cut, and expects the end mill be close to the size stated in its tool table, otherwise, during the roughing portion it will have a seriously wrong tool path.
You can look in fusion and see that compensation is supported in 2D contour, but not in 2D adaptive clearing. It is also supported in 2D pocketing, but it is only used on the finish pass, not the roughing passes. Also note that "enabling" compensation in Fusion means telling it to do it in the controller (on the machine in PathPilot via G41/G42) versus in the computer (internally by Fusion). Technically, you don't enable or disable compensation, you just tell Fusion where to do it. You will see this option on the "Passes" tab where it is available. It is available on most operations, however, as I pointed out above, actually only used on the "finish" portions of the operation.
One thing I learned about using CAM in Fusion (and I think this applies to all CAM) is to forget about hand coding. I fought it in the beginning, but now I think like Fusion does. You learn the operations and how they work (and their options) and then put them together in a right and efficient order to make the part. There are cases for hand coding an operation as still, but that is not about Fusion. For example, if I need to drill holes very accurately, by keeping all movements in the same direction (to deal with backlash), I might lay that out by hand (or rearrange what Fusion did), but for most (almost all) stuff, I think like Fusion, because that is what I am going to do it in. And I am happier now.