It's a hit rate thing, trying to make drafting more scientific and less art
Teams have various "measurables" thresholds at different positions that have to be met to spend a certain value of draft pick. Arm length for OTs, hand size for QBs, max throwing velocity for QBs, arm length for outside corner, short shuttle for ILB, short shuttle for S, three cone for edge, etc
They have data that shows players who meet these thresholds have a higher rate of "hitting" than those who don't, so if you filter out the players who don't meet those thresholds, theoretically your drafting should improve. Combine the improved hit rate with focusing on the more valuable positions early, and you should be doing much better than how teams were drafting in the 1990s, at the very least, although if every team is doing it to some degree, is there really an edge? Another way to think about it is these measurables are rare in humans, so draft the players with the rare measurables earlier, and then there should be more players available later in the draft with more common measurables to choose from, and missing later in the draft hurts you less
Some teams are all-in on this, while others are pretty lax about it
Regarding Banks, I don't see why he can't be Rashawn Slater, who's been a good NFL LT when healthy, and if he were in this draft with the knowledge of how he's played in the NFL (ignoring injuries), he'd probably go top five
Another thing to consider is interior OL (which apparently some teams view Banks and even Campbell as) are becoming more valuable because there are more dominant DLs and more difficult to deal with interior stunts than ever before, and the contracts for the top guards tracks with that