Skip to reading
The Amplified BriefDraft interactive brief
OverviewPolicy documents
Contents
Concept · Quality-adjusted productivity

Count the useful result—and the work required to get it

Explanation of policy measurement requirements

Compare useful results with the resources needed to produce them, including checking and correction. Faster output counts only to the extent that its value survives those demands.

A quick draft may be valuable. If it takes someone else longer to repair, the apparent saving can disappear.

When comparing workflows, ask:

  • Are we comparing similar work against a clear baseline?
  • Does the output meet the required quality standard?
  • What time, cost, checking and correction does the whole relevant workflow require?
  • Where is the gain demonstrated: a person's work, a team, or the entire service?

The framework separates efficiency, capacity, capability and creation. Those describe what changed. Amplification describes a material and disproportionate gain relative to resources. [§8.3] [Theory companion §12.1] [Measurement companion §22]

A simple numerical illustration

Hypothetical, not a result from the policy
Suppose a team previously completed 10 accepted estimates per week. With AI, it completes 30 equally demanding estimates that pass the same independent checks, using the same total weekly resources. The comparison includes preparation, checking, corrections and an allowance for setup and maintenance.

That is a threefold increase in quality-adjusted productivity in this simplified example. The estimates do not need to be a new kind of product for the gain to matter. A real claim would still need evidence that the comparison is credible and the gain persists.

Where to go next

Source passages

§8.3Policy framework: Evidence proportional to the decision§12.1Theory companion: Four productive results§22Measurement companion: Canonical outcomes and finding efficiency

Related

Examples

A limit worth testing

“This sounds like ordinary process improvement with an expensive evaluation layer.”

Optional. One short exercise with generated feedback. Nothing is scored or saved beyond this session.