Platform model

Judgment Line

The explicit boundary where AI may inform or draft but cannot own the call.

Canonical definition

The Judgment Line is the explicit boundary where AI may inform or draft but cannot own the call. It sits above tier labels (Delegate / Augment / Reserve) and below culture: if the line is unmarked, teams default to the Four Surrenders.

What decision failure does this prevent?

Executives want human in the loop without defining where the loop closes—dashboards drive meetings but no one states which decisions are human-owned.

Who commonly owns this problem

Executive sponsor defining human-only decisions before AI scales across the portfolio

Why it matters

Post-incident reviews debate the model, not the line the team crossed. Pairing the Judgment Line with Decision Rights tiers and CAN vs. SHOULD labeling turns vague governance into accountable architecture.

Practical example

A team asks whether they can enter a market; the model answers can with impressive completeness. Whether they should—given strategy, people, risk posture, and regulators—was never labeled as the question on the table.

What to do in the next 15 minutes

Pick one AI-assisted workflow and label each step: is the open question can or should? Mark where a named human must own the call regardless of model confidence.

Take the Leadership Assessment →

Journey phases

Next steps

When advisory applies: Engage advisory when tier labels exist on paper but incidents show the Judgment Line is still unmarked in production workflows.