I think there is something to the fact that drawing the diagram in front of someone - the performance of drawing - communicates something in itself. The way you place boxes initially to set out spatial relationships. The direction in which you draw the arrows. The speed at which you draw them - clearly and deliberately, or swiftly and as an aside. The things you drew and then had to erase and redraw bigger. The way you're able to augment the diagram as you go to elaborate on or clarify your audience's points of confusion. All of that is part of what you're communicating when you draw a diagram in front of someone.

None of that is part of what you capture in a static diagram in an architecture document.

That's one reason why I suspect that the diagrams in documentation shouldn't necessarily look like the ones you draw in a collaborative session.