Couldn’t you make some educated guesses and then stop when you arrive at a K that gives you meaningful clusters that are neither too high level nor too atomized.
Easier just to deliberately overshoot (with a too high k) and then merge any clusters with too much overlap.