The DMM7510 seems to uses some temperature measurement and numerical corrections (this can be seen during warm up). This can to a certain degree reduce the TC, if done for the different ranges. So they may not need ACAL to get the overall low TC. However this can not compensate for long time resistor drift. Having an automatic cal function could reduce the number of long time stable dividers / resistors and maybe reduce long time drift (e.g. beyond the usual cal cycle) if just a few, higher grade resistors are used.
Even if there is no explicit ACAL cycle, there could be some hidden, e.g. when changing rages. So the longer time it takes to change the function might be just a hidden partial ACAL function for the new selected function.
If done well, the ACAL way could give additional confidence, as drift of dividers could be noticed and it could work as a good and stringent internal self test. With sufficiently low noise of the ADC and amplifier ACAL can be relatively fast, especially if only 6 digit accuracy is aimed for. Usually the noise level is not that much lower in the high end meters, so they might need considerably longer to get the extra accuracy.
For a modern meter of this grade I would expect it to have some internal cal check and maybe adjustment. For a more normal use they might not need to make ACAL transparent to the user - for metrology use this would be requited.