Even the best SOPs are ignored. Not out of defiance, but because a procedure that lives beside the work is a procedure nobody reads at the moment of the decision.
The problem
Every organization has the same story. A procedure is written, agreed, circulated, and then quietly diverged from. Six months later the binder and the behaviour have nothing to do with each other, and nobody can say when they parted company. Adding AI makes this faster in both directions: more work gets done, and more of it drifts.
The usual response is training, reminders and audits. All three happen after the fact, and all three depend on someone remembering. That is why drift always comes back.
The mechanism
The compounding effect
Teams splinter into superstars who magnify their own lane and cannot lift the team, because what they know never leaves them. In APOP, completed work appends to the category's SOP as a matter of course. The playbook documents itself, the next person inherits it, and a new hire bootstraps from the same live context the AI uses.
Stale knowledge is the root cause of most AI failure, and most human failure too. This is where APOP attacks it.
That is what it means to live governed autonomy in a more organized, more reliable way. Not a mantra on the wall, but the default behaviour of every task in the queue.
A working session, not a pitch deck: your categories, your rules, and where governed autonomy would pay off first.