Kelvin probe force microscopy at normal pressure was performed by two different groups on the same Au-coated planar sample used to measure the Casimir interaction in a sphere-plane geometry. The obtained voltage distribution was used to calculate the separation dependence of the electrostatic pressure P res (D) in the configuration of the Casimir experiments. In the calculation it was assumed that the potential distribution in the sphere has the same statistical properties as the measured one, and that there are no correlation effects on the potential distributions due to the presence of the other surface. The result of this calculation, using the currently available knowledge, is that P res (D) does not explain the magnitude or the separation dependence of the difference Δ P (D) between the measured Casimir pressure and the one calculated using a Drude model for the electromagnetic response of Au. We discuss …