Measurement

Measurement is the layer that produces the evidence of return — not return as a generic concept, but return mapped to the canonical ROAI 4-Category framework: Productivity, Defensibility, Institutional, Category positioning. Its quarterly cadence is dictated by the time it takes for each category of return to surface meaningfully: Productivity in weeks, Defensibility at the audit cycle, Institutional across years, Category positioning at the strategic horizon.

Measurement is distinct from Optimization because Measurement is observational; Optimization is interventional. Measurement is distinct from Intelligence because Measurement reports on the function's own operations; Intelligence reports on the surrounding market and peer environment. Both Measurement and Intelligence inform the function's view of where it stands, but Measurement is the ledger of what the function has actually produced.

Measurement artefacts include the ROAI scorecard, the AI Bill of Materials cost ledger (DAT-06), the Adoption Metrics Dashboard (ADM-01), telemetry from deployed capabilities, and the Boardroom AI Reporting template (MAT-06) that translates operational metrics into the executive-stakeholder narrative. Measurement runs on instrumentation that Execution and Governance must have provided when capabilities were deployed; instrumentation cannot be retrofitted cleanly.

A function operating without Measurement has opinions about whether AI is working; it does not have evidence. The function may believe the AI portfolio is delivering productivity or risk reduction, but it cannot defend that belief to a CFO, audit committee, or board. The Measurement layer is the substrate on which the function's institutional credibility on AI rests.