Should chess engines visually show the reasoning behind their moves?

They have something called a PV, a principal variation. This is what leads an "island position".
What you would need is another engine that evaluates both engines island positions. As a chess player, we could then weigh certain variables higher or lower in importance.
Theoretically 2 engines now with the capability of a 3500 rating will seek drawn positions where human mortals will sacrifice pieces and play inferior positions to outwit an equally lower rated human.
We could use this second engine not to explain why but rather test out our thinking strategy. We would be the ones saying "I think I need to keep the bishop pair, let's see what Stockfish does with this restriction" or "I think this would be a good time to sack the exchange (with the same thing in mind)".
No....chess engines should work on the moves. But for the upmteenth time, there is software like https://decodechess.com/.