Vegetation Inspection Coverage Assurance Gap (VICAG)

GMI-EU-031Energy & UtilitiesIntegration-Required

Definition

Measures the percentage of vegetation-related outages or near-misses that occurred on spans where the AI/drone inspection had given a passing score, indicating the inspection method missed a hazard it should have detected.

The Measurement Gap

Traditional vegetation patrol and inspection metrics track patrol miles completed by cycle, equipment conditions observed by trained arborists, and cycle compliance rates. They evaluate whether the inspection program ran on schedule. They do not evaluate whether the inspection method (drone imagery, computer vision scoring) successfully detected all hazard types that subsequently caused outages.

The specific fall-short: drone and computer vision inspection is demonstrably effective at detecting vegetative encroachment. Branches entering clearance zones show clearly in aerial imagery. It is not designed to detect structural hazards: a mature oak with root rot that looks healthy from above, a species with known failure propensity under ice loading, or a leaning tree anchored by neighboring trees that will fall in the first high-wind event. An inspection that passes a span for encroachment but has no method for structural hazard detection creates coverage gaps that are entirely invisible in the inspection completion report.

The exact operational decision VICAG enables:

whether to supplement drone and computer vision inspection with ground-based structural assessment on spans rated highest-risk by the AI, and which hazard types the aerial inspection systematically misses.

What goes wrong without this information:

the utility invests in AI-enhanced aerial inspection and reports increased patrol coverage. Confidence in inspection completeness rises. The spans that receive passing scores are assumed clear. When vegetation outages occur on passed spans, they are attributed to "unexpected growth" or "storm conditions" without examining whether the inspection method was structurally incapable of detecting the hazard type involved.

How AI/ML makes this measurement possible:

vegetation analytics platforms record inspection scores by span with timestamps. OMS records outage events with span location and cause codes. Joining these records by span identifier (the computation VICAG requires) produces a post-inspection hazard miss rate that no pre-AI inspection method could have generated at scale.

Formula

VICAG = Veg_outages_on_passed_spans / Total_veg_outages

Variables:

Veg_outages_on_passed_spans: vegetation-caused outage events where the span had been inspected and scored as low-risk or no-action-required by the drone or computer vision system within the previous inspection cycle (count). Source: OMS vegetation-caused outage records joined to vegetation analytics platform inspection results by span identifier

Total_veg_outages: all vegetation-caused outage events in the measurement period (count). Source: OMS outage cause classification records

"Passed" defined as: the AI or drone inspection scored the span as low-risk, clear, or no action required, not merely that the span was inspected; a span inspected and flagged for action that subsequently had an outage is a different failure mode (execution gap, not inspection miss)

Calculation Frequency:

Per inspection cycle (quarterly or annually), with post-storm supplemental analysis for major vegetation outage events.

Edge Cases:

Include only outages where the cause was a pre-existing condition (hazard tree, structural compromise, species-prone failure) not new growth that occurred after the inspection. Storm-blown healthy trees are excluded; those are force majeure, not inspection misses. Separate encroachment-caused outages from structural-hazard-caused outages; VICAG should be reported by cause type to reveal which hazard categories the inspection method systematically misses.

Output:

Percentage. 0% = every vegetation outage occurred where the inspection had flagged concern. Below 10% = inspection provides meaningful hazard coverage. Above 25% = 1 in 4 vegetation outages occurred on spans the system passed. Action trigger: VICAG exceeding 20% for non-storm structural hazard outages indicates that ground-based supplemental assessment is needed on AI-identified high-risk spans.

Proposed formula, pending validation pilots.

At a Glance

Category
Vegetation Management Intelligence

Decision Supported

Is AI/drone vegetation inspection providing genuine coverage assurance, or is it creating false confidence by missing non-visual hazards (structural weakness, root compromise, species-specific failure propensity)?

← Trim Deferral Risk-Adjusted Savings (TDRAS)Vegetation Prioritization Execution Gap (VPEG) →

Assess Your Organization

The AI-GPS Assessment evaluates your AI governance maturity and recommends KPIs like this one for your implementation roadmap.

Take the Free AI-GPS Assessment