Agency & Ownership

What this dimension measures

This dimension looks at where decisions about new tools actually get made and how far a person has to travel to make one. When a team wants to try something low-risk, can it start on its own authority, or does the request go up and out before anything moves? When a trial is underway, is it clear who owns it, who can call it done, and who answers for the result? When a change touches more than one team, does it take one conversation or a season of coordination?

The assessment asks respondents about the most consequential decision concerning a new tool that their team can make without asking anyone else, and when that authority was last used. Both halves matter. Authority that exists on paper but has never been exercised tells a different story from authority used last month.

An organization with high Agency & Ownership tends to have clear owners within each team, low-risk trials that start without external approval, and cross-team changes that need a single ask rather than a committee. An organization sitting lower on this dimension may be full of interested, capable people whose ideas wait in queues for someone else's decision.

Whether those people also receive time, money, and help once they decide is a separate question, covered by Capacity & Commitment.

Why this matters for AI

AI opportunities tend to be local and perishable. The person who notices that a particular task could be done differently is usually close to that task, and the window in which the idea is worth pursuing is often short. If that person cannot act, or cannot find anyone who clearly can, the opportunity ages. By the time the decision reaches whoever has authority, the context has changed, the enthusiasm has cooled, or the idea has been overtaken.

This is why Agency matters more for AI than for a planned, centralized technology rollout. AI adoption is less like installing a system and more like hundreds of small decisions made by people who understand their own work. Each of those decisions needs an owner. When ownership is unclear, the decisions do not get made badly; they do not get made at all.

Ownership also governs what happens when something goes wrong. AI tools make mistakes, and someone has to decide quickly whether to adjust, bound, or stop their use. An effort with a clear owner can respond in a day. An effort owned by a committee waits for the next meeting.

Finally, AI work crosses boundaries. A change to how one team drafts its output changes what the next team receives. In an organization where each cross-team change needs several layers of coordination, these ripples become the main obstacle, and interested people learn to propose only what stays inside their own walls, which is rarely where the value is.

How this can work in your favor

When your responses describe clear owners and low-risk trials that start without outside approval, AI adoption can move at the speed of the people who see the opportunities.

Ideas get tested while they are still relevant. Someone who notices that a task could be done differently can try it this month, not after a planning cycle. The opportunity is tested while the context that produced it still holds.

Responsibility is clear when it counts. When an AI tool does something unexpected, there is a named person who can decide what to do, and who is expected to. Problems get addressed rather than escalated into silence.

Interest turns into action. Enthusiasm in an organization is a resource that decays if it cannot be used. Where people can act, their interest becomes trials, and trials become evidence. Where they cannot, it becomes frustration.

Cross-team changes are possible. If a change that touches two teams needs one conversation rather than a committee, the organization can pursue the AI uses that cross boundaries, which are often the ones worth having.

Many small bets replace one large one. Because decisions are distributed, the organization learns from many modest efforts rather than staking everything on a single centrally chosen project.

How this can make AI adoption harder

When your responses suggest that owners are unclear and that even low-risk trials need approval from outside the team, AI adoption tends to stall before it starts.

Interest that cannot become action. The people who see opportunities are not the people who can authorize pursuing them, and the path between them is long. Ideas wait in coordination cycles until the moment passes. In time, the people closest to the work stop proposing.

Nobody owns the result. When a trial does begin, it is unclear who decides whether it worked. It runs, people have opinions, and no one has the standing to call it. Efforts linger rather than conclude.

Slow response to problems. When an AI tool produces something wrong, the question of who may act on it goes up the same long path. Meanwhile the tool keeps being used, or is quietly abandoned, without a decision either way.

Work confined to silos. Because cross-team changes need many layers of agreement, people propose only what stays inside their own boundary. The AI uses that would matter most usually cross boundaries, so they are never proposed.

Enthusiasm at the top, paralysis below. Leadership can be genuinely ambitious about AI while the organization produces little, because the ambition came without authority for anyone to act on it.

This is friction, not a lack of capability. People in such an environment are often entirely able to adopt AI; the dimension measures whether they are allowed to.

What this dimension is not

Agency & Ownership is not the absence of oversight. An organization can have clear limits on what teams may decide and still sit high on this dimension, as long as teams can act within those limits without assembling permission each time.

It is not Strategic Ambition. Leaders may want AI very much while the people who could make it happen lack any authority to try; the two dimensions are separate for exactly this reason.

It is not the same as resourcing. Being allowed to start a trial and being given the time and money to run it are different conditions. Capacity & Commitment covers the second.

It is also not individual heroics. A single person who gets things done by working around the organization is evidence of low agency, not high.

What we look for

The assessment looks for decisions that were actually made within a team, not for statements about empowerment. Signals that tend to be informative include:

• The most consequential decision about a new tool that a team can make without asking anyone else, and when that authority was last used.
• Whether a recent low-risk trial started on the team's own say-so or waited for approval from outside.
• Whether a person can get permission and modest resources to test an idea without a formal proposal.
• Whether a recent effort had a named owner who could call it finished, or whether its status depended on who was asked.
• How many separate agreements were needed for the last change that touched more than one team.
• Whether good ideas from people close to the work have reached a decision, or are still waiting.

"We empower our people" is not evidence. The name of the person who decided, and how long it took them, is.