Definition
The change over time in mean-time-to-resolve for novel supply chain situations (counterfeit recognition, vendor disputes, complex substitution validation, unusual receipt discrepancies) that automation does not handle, measured against a tenure-matched baseline from the pre-automation period.
The Measurement Gap
Legacy MRO supply chain metrics do not measure whether the human workforce is losing capability on novel cases as automation absorbs routine work. Cycle time measures current performance against current volumes. Productivity measures throughput. Quality metrics are dominated by routine work the AI handles. When AI receipt platforms, automated vendor-dispute workflows, and AI substitution recommenders absorb the routine cases, the storeroom and procurement teams handle only the genuinely hard cases the AI flags as exceptions. But those teams no longer practice on the routine cases that built their judgment. Five years of declining novel-case volume means five years of declining practice on the judgment that novel cases require. The degradation shows up only when a high-profile novel case fails badly, and then the failure is attributed to the case rather than to the structural atrophy that produced an ill-prepared response. The decision SWAI supports: whether to invest in cross-training, simulation-based skill refresh, or human-in-the-loop policy for novel cases, and whether automation has progressed past the point where humans can intervene effectively.
Formula
SWAI = MTTR_novel_current / MTTR_novel_baseline (matched by tenure)
Variables:
- MTTR_novel = mean-time-to-resolve for cases flagged as novel or exception within a stable taxonomy
- baseline = same team or matched-tenure cohort's MTTR at the start of meaningful automation deployment
- Tenure-matching prevents new hires (who naturally take longer regardless of automation) from dragging the number
- Use moving 12-month window
Worked example: Baseline period (pre-automation): storeroom team handled novel discrepancy cases in MTTR of 3.2 hours. Current period: same team, tenure-matched, handles the same case-type taxonomy in MTTR of 5.8 hours. SWAI = 5.8 / 3.2 = 1.81. Eighty-one percent increase in resolution time, a strong atrophy signal. By case type: novel counterfeit triage SWAI = 2.4 (severe); vendor dispute SWAI = 1.3 (mild); cross-plant substitution adjudication SWAI = 1.8.
Edge cases: Novel-case taxonomy must be stable over the comparison window. Tenure-matching is essential: new hires inflate raw MTTR. Some atrophy is appropriate (humans should not maintain expert performance on cases they rarely see); the question is whether the rate exceeds operational tolerance.
Data sources: Exception case events with type, accept-time, resolution-time, and resolver from procurement and storeroom case logs. Staff tenure from HR system. Automation deployment timeline from project records.
Confidence Grading (required alongside every SWAI value):
| Grade | Basis | Interpretation |
|---|---|---|
| A: Empirical | 5+ observed events in measurement window provide direct tail observation | Measurement grounded in observed outcomes; highest confidence |
| B: Hybrid | 1-4 observed events plus Monte Carlo simulation with Bayesian updating | Blends observation and simulation; moderate confidence |
| C: Simulated | Zero observed events; derived entirely from Monte Carlo simulation | Modeled, not observed; treat as indicative, not validated |
A SWAI of any value at Grade C is a simulation that agrees with itself. It tells you the model is internally consistent but not that it matches reality. Grade C scores should trigger proactive stress-testing against historical near-miss events and cross-asset analogues rather than acceptance as calibrated.
Proposed formula, pending validation pilots.