For a windshield or eye glasses... why not try it on something else first. I find it hard to believe that the layer will be uniform or bring the reflectivity at that surface to 0.6%. No harm in trying on specimens. The transmitted wavefront will pick up the NΔT distribution over the surface; in reflection, double that. I was a vacuum coating operator/tech/engineer/scientist for 12 years. Never worked the chemical deposition stuff though. My boss, who was a chemist, dabbled in those things... but to no avail.
OH! You would have to be able to strip off the old coating, if present. Otherwise, the liquid stuff will increase the reflectivity by ~frustrating~ the existing coating. I once made that mistake... with nearly disastrous results, before figuring out what was going on. I'm guessing if not applied properly --- might wind up looking like a soap bubble or oil slick. I actually researched this phenomenon to determine/measure non-contact the thickness of the tear layer on the human eye cornea. The context was Bausch & Lomb Eye Drops and contact lenses development. Pleases me that to this day, one of B&L's most profitable products is --- OTC and Prescription Eye Drops! Tom