There is non-manual ways to do it (called self-calibrating), but those need many more antennas. (What you really need is good coverage of the (u,v) plane AND many more different baselines between pairs of telescopes than the number of individual antennas.) If you do not have that self-calibration fails and leads to horrible image artifact or does not converge at all. They limited the influence of the manual calibration by observing a Quasar which is basically a point source between subsequent observations of M87* to get an independent amplitude calibration.