Explainable score
Five bounded maintenance signals with a visible contribution beside every result.
Revolution Road: Nocturne
Know what needs attention.
Keep work moving.
Revolution Road connects assets, work orders, inventory, stores, districts, and accountable maintenance activity in one operating workspace.
Phase 9 release
Start simple. Earn sophistication.
The transparent baseline is useful today and establishes the clean feature history, labeled outcomes, and feedback a supervised model needs tomorrow.
Human decision
Accept, reject, or defer a suggestion without silently changing the maintenance plan.
Measured promotion
Backtests, drift evidence, artifact checksums, and shadow mode before supervised activation.
From field evidence to action
One accountable chain.
Daily ETL creates a privacy-filtered feature snapshot. The active model scores it. Revolution Road publishes reasons and confidence. A person decides. The outcome feeds the next evaluation window.
See every Phase 9 deliverable →ObserveUsage · failures · PM · sensors
ExplainScore · confidence · factors
DecideAccept · reject · defer
LearnBacktest · drift · retrain review
August 2026
Predictive maintenance, without the black box.
The risk register, model card, privacy-filtered analytics pipeline, feedback loop, validation evidence, and drift controls are ready. Supervised time-to-failure stays in shadow until connected operational history proves it.
Operations workflow update
Store requested. District delivered.
Supplies Needed closes the handoff with scoped store requests, a district inbox, partial quantities, append-only delivery evidence, and CSV/API access.
Team and Market Structure
Snow Star LP, from market to store.
A navigable Market → District → Store hierarchy now places Market Presidents, District Managers, and General Managers at the right operating scope with preserved assignment history.
Methodology whitepaper