The data on spectroscopic galaxy clustering collected by the Dark Energy Spectroscopic Instrument (DESI) will allow the significant detection of subtle features in the galaxy two-point correlation in redshift space, beyond the ‘standard’ redshift-space distortions. Here, we present an independent assessment of the detectability of the relativistic dipole in the cross-correlation of two populations of galaxies if they would be selected from the Bright Galaxy Survey (BGS) of DESI. We build synthetic galaxy catalogues with the characteristics of the BGS using the light cone of a relativistic N-body simulation. Exploring different ways of splitting the populations of galaxies we find that with an unequal split with more bright galaxies than faint galaxies the detectability is significantly boosted, reaching 19σ in the redshift bin 0.2 ≲ z ≲ 0.3 and expected to be even higher at lower redshift. Moreover, we find that the measured dipole agrees very well with the prediction of relativistic effects from linear theory down to separations of ∼ 30 Mpc h^−1.