Inside an enterprise AI rollout, week one.

Day one opens with the people, not the model. The first session is ninety minutes and the AI barely comes up; the team discusses what each person does now, what stays theirs, and what the system will take off their plate. "Augment, never replace" is described not as a slogan but as the literal agenda of day one.

By midweek the team laminates a single-page cheat sheet for every role, listing what the AI can do, what it must never do alone, and the exact phrase to use to request a human check. People keep them, taped to monitors months later. A laminated card outlasts a slide deck because you can hold it at 2pm when something feels off.

Three components: "what it does" (the three tasks the system handles for that role), "what it never does alone" (the decisions that always route to a person, in plain language), and "the escalation phrase" (one sentence that flags a case for human review). It teaches people exactly where not to trust the AI.

A runbook is the spine of every workflow the AI touches: the normal path, the edge cases, and the rollback if output looks wrong. The team walks each group through their runbook by breaking it on purpose, feeding the system a bad input and watching the human catch it, so the catch is rehearsed before it's needed.

It is not accuracy. The week closes on one number: how many people in the room can say, without checking the card, where the human stays in the loop. The post says that's the only week-one result Silicon Prime cares about, since the model improves over months but the habits must hold from the start.

The post is explicit: nobody lost a job that week. Everybody got a card, a runbook, and a clear line on where the human stays in the loop. The framing is "augment, never replace," with week one spent teaching people exactly where not to trust the AI rather than displacing them.

The post argues week one isn't about teaching people to trust the AI, it's about teaching them exactly where not to. The model will improve over months, but the habits, knowing where the human stays in the loop, are the thing that has to hold for the program to still be running a year later.

Because a responsible rollout spends week one on people, cards, runbooks, and rehearsed human catches rather than rushing the model into production. The post frames that deliberate pace as the reason the program is far more likely to still be running a year from now.