How do we measure uncertainty for a reflectometer???

We got a nonconformance in our recent pre-A2LA audit for not having measurement uncertainty addressed for our reflectometers. We have glass standards (a mirror) in place that we use to calibrate (verify) our reflectometers.

The glass standards are N.I.S.T traceable, They are measured on the "Hoffman" light meter / sphere. Basically a reflectometer set up to the SAE J964 method.

This light meter and light source are both sent out for calibration along with the detector once a year (they have NIST traceablility then).

Does anyone know how to address this? We want to comply with A2LA requirements but this is the ONLY kind of measurement item that we thing that we can't meet the standard to.