Design an AI/ML roadmap covering predictive maintenance for [AssetClasses] with current MTBF of [MeanTimeBetweenFailure] and demand forecasting for volatile volumes ([DemandVolatility]). Catalog available [DataSources], define model candidates, and stage experiments that reduce unplanned downtime and stockouts while optimizing [SparesLeadTime]. Align delivery to value within [TimeHorizon].
Output format:
Data Readiness Assessment (table: Source | Quality | Completeness | Latency | Owner).
Model Options (table: Use Case | Model Type | Features | Expected Lift | Risks).
Experiment Plan (table: Pilot | Site/Asset | Hypothesis | KPI | Stop/Go Criteria).
MLOps Design (table: Stage | Tooling | SLA | Monitoring | Responsible).
Benefits Case (table: KPI | Baseline | Target | Dollar Impact | Realization Timing).
Address human-in-the-loop, maintenance work orders integration, and planner adoption.