Give the grid eyes.
Autonomous inspection data, thermal imagery, LiDAR measurements and AI findings become condition scores, prioritized work orders and GIS-ready evidence.
This screen uses labeled demonstration data. Flight control, GIS writes and utility work-order dispatch remain approval-gated until live integrations are connected.
Mission control
AI inspection engine
The first production slice accepts field measurements and returns a risk score, severity, evidence reasons and recommended action. Computer-vision models plug into this same contract later.
Prioritized findings
| Asset | Finding | Severity | Confidence | Condition risk | Recommended action |
|---|---|---|---|---|---|
| TX-09-441 Transformer | Thermal hotspot +23.6°C above baseline | High | 97% | 84 | Crew inspection within 24 hours |
| PL-44192 Distribution pole | Crossarm deterioration and 4.8° lean | Medium | 93% | 63 | Field inspection within 14 days |
| INS-44192-B Insulator | Surface crack candidate | High | 91% | 78 | Confirm at next safe access window |
| ROW-17-88 Right of way | Vegetation clearance 7.2 ft | Medium | 99% | 61 | Vegetation work order |
| SW-18-02 Switch | Minor corrosion pattern | Low | 86% | 34 | Track and compare next flight |
Thermal intelligence
Compare component temperature against asset-specific and fleet baselines. Flag connectors, transformers, switches and other abnormal heat signatures.
LiDAR + vegetation
Measure conductor clearance, encroachment, pole geometry and sag. Turn distance thresholds into prioritized vegetation work.
Repair verification
Re-fly completed work, compare before/after evidence and close the maintenance loop with a time-stamped audit trail.
Integration backbone
Mission metadata, telemetry, RGB frames, thermal frames and LiDAR files.
Findings, confidence, condition score, evidence and trend history by asset ID.
Read asset geometry and IDs; write approved condition attributes, attachments or related records.
Create maintenance recommendations only after the utility's configured human approval step.
Pilot hardware package
Inspect → detect → prioritize → repair → verify.
The software slice is structured for real drone feeds and utility integrations. Next step is connecting one pilot drone, one corridor and one utility asset export.
