My guess is if you observed a highly enough experienced and successful dowser for long enough you'd find out diagnostic work depends on a lot more unconscious knowledge than most give it credit. As an engineer I know to look for a slight depression in the ground, maybe so slight most can't notice, and the smell of wet dirt is different, and pipe comes in coils or segments of length X so even unconsciously I might eventually learn to detect at spots multiples of X away from the water main. I imagine I could unconsciously become a highly skilled dowser just by working it hard enough for enough years and storing enough data.
I know for a fact I can repair old ham radio equipment this way. I can "smell", or something, failed electrolytics far better than using an ESR meter by hand or replacing randomly. Only takes 40 years of experience and you can learn it too. I'm not marketing my skill as a form of witchcraft I'm well aware its merely decades of unconscious study. It is not even remotely controversial that different humans have different abilities to focus their eyes and resolve small things and see in darker environments; I think it quite likely indeed that I can unconsciously see bulging capacitor packages better than the average bear. I know for a fact, having replaced many, that old HP/Tek instrument-grade "black beauty" capacitors have a, shall we say, unique smell when soldered as part of the removal process, and I feel unconsciously that I can smell a cracked/leaking "black beauty" capacitor at room temperature whereas most people cannot. I do not smoke, for example, so I know for a fact my sniffer works better than anyone who smokes anything, but I also think that having screwed around with electronics for several decades I can quite literally smell a failed "black beauty" capacitor (edited: at room temperature; I assure you anyone can smell it when overheated LOL).