Seven measurable elements form a closed loop. Human Agency isn't one of them — it's the halo around the whole loop, read from how tightly the moderating relationships hold and whether feed-forward stays earned or becomes a shortcut around them.
Not a node. It's the diagnosis read off the loop's pattern: tight moderation with an unused or absent feed-forward arrow reads as agency preserved; loose moderation with a strong, confidence-driven feed-forward reads as agency quietly eroding.
How much autonomy AI actually has, and how much is actually handed off in practice — the governed/dependent side of the loop.
The objective, measurable results of AI-enabled work — independent of how anyone perceives them.
Leadership's belief about whether AI is "working" — can diverge from the outcomes themselves, in either direction.
The actual structural action an org takes — charters, policy, formal decision-rights — that closes the loop by acting on Tech Agenticness & Delegation. This is what "governs," in practice.
A person's own attitude, resolved from behavioral signals — moderates their personal delegation behavior.
The aggregate sentiment mix across a team — moderates delegation through norms and peer effect.
The org-wide climate toward AI — now a tributary moderator alongside Individual and Team Sentiment, distinct from the formal Governance Response it may or may not translate into.