Definition
The fraction of counterfeit detection modes specified by applicable standards (AS5553, AS6081, OEM-specific) that the AI-augmented receipt inspection actually detects with validated sensitivity, surfacing whether a confident-looking detection rate is high on common modes but zero on sophisticated ones.
The Measurement Gap
Legacy MRO supply chain metrics measure AI counterfeit detection at the outcome level without examining mode coverage. AI detection rate measures how often the AI flags counterfeits among inspected items, but only on modes the AI covers. Receiving inspection cycle time measures speed. Rejection rate measures volume. None of these tells you how many of the standard's required counterfeit detection modes the AI actually addresses. Counterfeit detection standards enumerate specific failure modes: visual marking forgery, surface finish anomalies, internal void inspection via X-ray, material composition mismatch, lot-traceability gap, provenance-chain break. The AI may excel at visual and surface modes while covering zero internal-construction or material-composition modes. A 96% confidence on modes covered tells you nothing about modes not covered. Without mode-level coverage measurement, the operation accepts a partial detection capability as if it were a complete one, and the gap surfaces as a counterfeit acceptance event on an uncovered mode. The decision CMCS supports: whether AI-augmented counterfeit detection can replace traditional inspection on a part class, what supplemental inspections are required by criticality, and whether OEM certification gate is needed.
Formula
CMCS = detectable_modes_by_AI / required_modes_per_standard
Variables:
- detectable_modes_by_AI = modes with independently-validated AI detection sensitivity at or above the regulatory threshold
- required_modes_per_standard = modes enumerated in the applicable standard for the relevant part class
Compute per part class and per applicable standard separately.
Worked example: Aerospace fasteners under AS5553 require detection of six modes. External marking forgery, surface finish anomalies, internal void inspection, material composition mismatch, lot-traceability gap, provenance-chain break. AI vendor's validated capabilities cover external marking, surface finish, and provenance-chain (limited). CMCS = 3 / 6 = 0.50. Without supplemental manual or laboratory inspection on uncovered modes, the AI alone leaves half the standard unaddressed.
Edge cases: Standards differ by part class. Aerospace AS5553, electronics AS6081, defense DFARS counterfeit clauses, nuclear NQA-1. Compute per applicable standard and per class. Validated sensitivity must come from independent testing, not vendor self-reporting. For partially-covered modes (AI catches some but not all instances within a mode), apportion credit conservatively: 0.5 for partial, 1.0 for full.
Data sources: AI detection capability from vendor validation testing (independently verified). Standard mode enumeration from applicable standards documentation. Part-class-to-standard mapping from item-master regulatory tagging.
Proposed formula, pending validation pilots.