Dimension 01

Capture

The root dimension

Whether expertise, judgment, and edge-case knowledge get written down and structured — or stay locked in the heads of whoever happens to hold them.

Level	What Capture looks like
AQ1 — Individual	Knowledge exists only in individual heads and private notes. Nothing is structured for reuse.
AQ2 — Team	Some knowledge is written down informally — shared docs, prompt libraries — but scattered and undiscoverable outside the team.
AQ3 — Connected	Knowledge is captured with enough structure that systems, not just people, can reference it across team boundaries.
AQ4 — Operational	Capture is a deliberate, owned practice. Judgment that required a specific person is written down well enough for someone else — or an agent — to execute.
AQ5 — Adaptive	Capture is continuous. Every execution adds to the store automatically; the organization never has to restart knowledge capture because it never stopped.

Dimension 02

Operationalization

Whether captured knowledge has become a repeatable workflow or decision rule that doesn't depend on any one person being in the room.

Level	What Operationalization looks like
AQ1	No workflows exist beyond what's in one person's head.
AQ2	Workflows are copy-pasted between individuals; nothing is built into a system.
AQ3	Workflows are defined well enough that agents can execute steps without a human re-deriving the process each time.
AQ4	Workflows are the operating playbook — versioned, owned, and structured for human and agent execution.
AQ5	Workflows update as the business changes because operationalization is wired to the feedback loop, not frozen at time of writing.

Dimension 03

Amplification

Whether people and AI agents are executing the operationalized system together, at real scale — versus the system existing but sitting mostly unused.

Level	What Amplification looks like
AQ1	A handful of individuals get personal leverage from AI. Nobody else does.
AQ2	Small groups get leverage from shared patterns, but reach is capped at people who happen to know.
AQ3	Agents execute across functional boundaries — the leverage is organizational, not personal, for the first time.
AQ4	Agents execute the actual operating playbook directly, at the scale of the whole function or company.
AQ5	Amplification compounds: this month's execution makes next month's execution better through feedback.

Dimension 04

Feedback & Improvement

Whether the results of human and agent execution flow back in and sharpen the system, or whether captured knowledge slowly goes stale.

Level	What Feedback looks like
AQ1	None. Whatever an individual learns stays with that individual.
AQ2	Informal — someone occasionally updates a shared document when they notice it's wrong.
AQ3	Ad hoc feedback exists but isn't structured or owned; it depends on someone noticing and caring enough to fix it.
AQ4	The playbook is reviewed on a cycle — quarterly, after incidents, or at another owned interval.
AQ5	Every execution generates signal that updates the system continuously and structurally.

Dimension 05

Governance & Trust

Whether there is a defined, documented, and named path for validating agent output, catching errors, and intervening before a decision becomes irreversible.

Level	What Governance looks like
AQ1	None needed yet — individual use, individual risk, nothing shared to govern.
AQ2	Informal peer review within a team; no defined ownership.
AQ3	Cross-boundary execution exists, often without a matching cross-boundary validation path.
AQ4	Validation, escalation, and rollback paths are documented and owned, matching the scope of agent execution.
AQ5	Near-misses and overrides are captured and folded back into the playbook through the feedback loop.