Why does adding automation not fix a broken system?#
Automation amplifies whatever system it lands in. Where intent is clear and results are checked, it speeds things up. Where they are not, it produces unchecked work faster — which is worse than slow, because the cost arrives later and lands on someone else.
How do you tell substitution from change?#
By what moved. If the tools changed and the handoffs, approvals and decision rights did not, nothing structural has happened. The constraint is still where it was, and the new tool has made the queue into it shorter — see constraint migration.
This is the mechanism behind most of the AI productivity paradox: real individual gains that never reach delivery, because the system they landed in was never the thing that was slow.