Hold the ceiling at A2 for strategic workforce planning. Planning is judgment work. Agents model and author the position; the headcount envelope is a board-level commitment that a human must own. Raise it only after two consecutive quarters where the override rate stays under five percent and every override has a written cause.

Keep the repeatable plan scenario volume in the spine and pull the exception tail back into the function. The spine is cheaper per unit; the tail is where the context that prevents rework actually sits.

Hold the ceiling at A3 for talent acquisition. Sourcing, screening and scheduling reach A4. The hiring decision itself never does — it stays a named human decision with adverse-impact evidence attached. Raise it only after two consecutive quarters where the override rate stays under five percent and every override has a written cause.

Keep the repeatable requisition volume in the spine and pull the exception tail back into the function. The spine is cheaper per unit; the tail is where the context that prevents rework actually sits.

Hold the ceiling at A4 for onboarding. The highest-autonomy people process in the enterprise. It is deterministic, well-instrumented, and every exception has a named owner. Raise it only after two consecutive quarters where the override rate stays under five percent and every override has a written cause.

Hold the ceiling at A2 for hr business partnering. Tier-1 policy questions are fully agentic. Anything with disciplinary, exit or litigation potential is human-decided by design — the agent only prepares the pack. Raise it only after two consecutive quarters where the override rate stays under five percent and every override has a written cause.

Agents in this lane may act without a human when the plan scenario sits inside the stated value, risk and confidence envelope. Outside it, the unit holds at a gate with a named approver and a running clock.

Every autonomous decision writes inputs, the rule version applied, the model and prompt version, the output and a reversal path. An action with no evidence record is treated as a control failure, not a fast decision.

Escalation is mandatory when confidence falls below the floor, when two options sit inside the confidence band, or when a break condition on a ratified position fires. VP Workforce Strategy owns the response clock.

The ceiling for this lane is 62 on the platform scale. Planning is judgment work. Agents model and author the position; the headcount envelope is a board-level commitment that a human must own. The ceiling is a governance decision, not a technical limit, and only VP Workforce Strategy can propose moving it.

Agents in this lane may act without a human when the requisition sits inside the stated value, risk and confidence envelope. Outside it, the unit holds at a gate with a named approver and a running clock.

Every autonomous decision writes inputs, the rule version applied, the model and prompt version, the output and a reversal path. An action with no evidence record is treated as a control failure, not a fast decision.