You're right that chirality is important and this thing isn't able to probe that. As to what I'll broadly call "sample prep", you're right again to point out how that is a critical consideration. So you're double right. Operating in NIR does give deeper penetration, which is one of the reasons to consider NIR spectroscopy in an application; I'd guess that this thing operates somewhere in the 700–900nm range, but I admittedly don't know what kind of interrogation volume that gives for a pill.

So, I'm wrong to casually dismiss opacity of the target. That stems from a bias due to a background in instrumentation (meaning, I start my thinking there). I think I'm still right though to think that instrumentation is a bigger problem than sample prep for this device, which I didn't do a good job in stating earlier.