Agree with all architects, we use digital levels to check greens, mostly to assure a minimum 1.5% slope for drainage, 3% (actually up to 3.5% for me) maximum slope in cupping areas, and max <10% for any rolling edges (much steeper and they dry out) and <20% for any sharp tiers, to avoid mowing scalp.
Doing so many public courses, and having a design philosophy of (mostly) gently rolling greens proportionally challenging putts the further away from the hole that they are, its hard to say than any slope between the 1.5-3% range is "wrong". And that's even when I plan my basic slope swales to be different on every green to avoid a "standard break". (i.e., drawing some at 1.75%, some at 2%, some at 2.25%, etc.)
If an architect puts a small, random hump in somewhere, that could easily get lost in topdressing (or construction shaping). But even then, you could argue that a less prominent mid green hump might be a more effective hazard/challenge as it is less visible to the eye, no? I'm not sure there would be any "correct" height or size to those.
And, as mentioned, even if I had an intention to deflect a putt from one part of the green via subtle contour to another, what is the random chance of a combo of approach shot finish location on a day when the pin in is where it needs to be?