The extended main-sequence turn-offs (eMSTOs) and extended red clumps (eRCs) observed in intermediate-age star clusters challenge the traditional understanding of clusters as simple stellar populations. Recently, eMSTOs have been interpreted as signatures of stellar rotation. In this work, we test the effectiveness of rotational mixing in shaping the colour-magnitude diagram (CMD) of star clusters. We computed a set of separate single-age synthetic stellar populations, referred to as ‘Base Stellar Populations’ (BSPs), including stellar rotation. These BSPs were generated from two grids of stellar models that share the same input physics but differ in the efficiency of rotational mixing. We used an optimization algorithm to determine the best combination of BSPs to fit the CMDs of two star clusters: the Small Magellanic Cloud cluster NGC 419 and the Milky Way cluster NGC 1817. The synthetic clusters with weak rotational mixing provide the best fit to both the eMSTO and eRC features for both clusters, and are consistent with the luminosities and asteroseismic masses we derived for eRC stars in NGC 1817. In contrast, synthetic clusters with strong rotational mixing result in overly bright post-main-sequence stars, inconsistent with observations. This suggests that, for intermediate-mass stars, the influence of rotational mixing of chemical elements on stellar evolution cannot be so strong as to significantly increase the post–main-sequence luminosity. A simple test suggests that accounting for self-extinction by decretion discs in equator-on fast rotators could influence inferred rotation distributions and help reconcile the projected rotational velocity discrepancy across the eMSTO between models and observations.