You can use the Karl Fisher method, but it isn't as easy as Wikipedia describes. The hardest part is definitely not purchasing anhydrous methanol. You have to calibrate the system carefully, and execute the protocol very carefully. With practice, you can get good at the procedure, and when the rig is set up properly, the system can be flawless. The darn problem is that, if you're a human, working on the surface of the planet Earth, there are always traces of moisture. And if the sample gets contaminated, the auto-dispensing titrator just goes haywire. And you have to let it finish before you flush out the chamber and start over.