Local-first pilot path
GridAPM is positioned for controlled transformer APM pilots that can start from approved datasets and offline-capable review workflows.
Security and deployment
Local-first controls for transformer APM pilots
GridAPM is designed for utility environments where transformer evidence, AI assistance, reports, and approvals need clear boundaries before broad deployment.
Security posture
Built around controlled evaluation, not open-ended automation
A GridAPM pilot should define what data is approved, what AI may assist with, who reviews output, and which decisions remain outside the software boundary.
Engineer approval
AI-assisted recommendations remain draft material until responsible engineers review, edit, approve, reject, or escalate them.
Evidence traceability
Material findings should preserve source evidence, assumptions, policy context, reviewer state, and audit history.
No autonomous control
GridAPM public materials do not claim autonomous transformer protection, switching, control, or final operational authority.
Deployment boundaries
Hosted services, telemetry, and external AI providers should be enabled only when a deployment profile explicitly permits them.
Utility OT respect
Pilot scope should respect utility security reviews, operational technology segmentation, sensitive data boundaries, and procurement controls.
Pilot controls
What security and engineering teams should define first
Security review is strongest when the pilot starts narrow: approved evidence, known reviewers, local workbench expectations, and clear deployment boundaries.
Approved transformer population and evidence scope
Local evidence files, source provenance, and review owners
Defined AI advisory boundaries and human approval gates
Pilot-specific data handling, retention, and export expectations
Security review before enterprise integrations or hosted services
Clear separation between draft recommendations and approved actions