The coherent photonuclear production of a $mathrm{J/}psi$ vector meson at the LHC has been computed using two different sets of solutions of the impact-parameter dependent Balitsky-Kovchegov equation. The nuclear dipole scattering amplitudes are obtained either from ($i$) solutions for this process off proton targets coupled with a Glauber-Gribov prescription, or ($ii$) from solutions obtained with an initial condition representing the nucleus. These approaches predict different cross sections, which are compared with existing data from ultra-peripheral collisions at the LHC. The latter approach seems to better describe current measurements. Future LHC data should be precise enough to select one of the two approaches as the correct one.