Read the inspection record, detect against the criteria your engineers configured, bind every finding to the frame it came from, and hold each one for the named engineer who dispositions it.
Ingest inspection imagery, LiDAR passes and patrol records from supported asset and inspection systems.
02
Normalise frame, pass and asset identifiers, and carry each capture forward with the record it came from.
Reason
03
Detect against the criteria configured for the asset class, and score each finding.
04
Apply the utility's own severity bands, clearance rules and asset-class thresholds.
05
Bind every detection to the frame, the structure and the inspection it came from.
Decide
06
Mark what the imagery did not cover — the angle not flown, the span obscured, the pass cut short.
07
Route every detection to the named engineer who dispositions it.
Out
08
Retain the imagery, the detection, the engineer's disposition and the outcome against the asset.
09
Execute write actions only inside the approval boundaries agreed during implementation.
→Product statement
The agent proposes detections; a named engineer dispositions each one, and the maintenance programme — not the detection — is what the auditor asks to see.
Example workflow
One detection, frame to disposition
AgentHuman
1Inspection record receivedAerial pass, ground patrol, thermal scan or LiDAR survey
2Asset context assembledStructure, circuit, asset class, prior findings and capture conditions, each with its source
3Detection proposedAsset, defect class, frame and confidence
4Controls appliedCriteria checks, coverage checks, prior-finding checks and confidence threshold
No human action required
Stages 1 to 4 run unaided, and nothing is dispositioned at any of them — the agent is detecting, and the engineer's lane opens at the confidence gate.
5DecisionBranches at the confidence threshold
High confidence
Goes to the named engineer to disposition.
Low confidence
Adds a second-reader read first.
Engineer disposition
The detection is held with its frame, its coverage gaps and the confidence.
Confirm · Reclassify · Send to second reader
Dispositioned — sent to the programme▼
6Asset systems updatedOnly where write access and approval policy allow it
7Outcome evaluatedCorrection rates, coverage gaps, field verification and defects surfaced later
Corrections
Every engineer correction is counted in the evaluation.
What should not run autonomously
Human approval stays in control
Outside the boundary — human approval required8 items
Dispositioning a defect, or clearing one from the list.
Changing an inspection interval, or the coverage flown.
Declaring an asset, a span or a structure healthy.
Scheduling, deferring or cancelling maintenance work.
Automation boundaryAgent acts unaided
✓Read the inspection record against the configured criteria.
✓Bind each detection to the frame and the asset it came from.
✓Carry the pass, the imagery and the capture conditions.
✓Mark what the imagery did not cover, and hold for an engineer.
Any write happens inside the boundaries agreed at implementation, never ahead of disposition.
Closing a maintenance record or an unresolved issue.
Recording that an annual inspection was performed.
Deciding that a span met its clearance distance.
Changes to detection criteria or severity rules.
Example output
One detection, annotated
A detection rests on a frame, and the frame is the whole of what the agent saw.
Detection output · single assetIllustrative example
Asset
Detection
Frame
Source of record
Confidence
Held for
Lattice suspension tower
Corrosion at a suspension-clamp fitting, tower body clear
AR-0417
Aerial pass on file
88%
Named engineer, not dispositioned
As receivedTaken from the inspection pass and the asset register — nothing on this side is written by the agent.
Small fittings are the least-counted things on the line — hardest to resolve, hardest to label, and they fail without warning. Nestack reports the engineer-correction rate by slice, not only in total.
Slice performance — reported separately, not only in aggregateIllustrative example
Slice
Failure rate
Lift
Lift vs. threshold
Status
Small fittings and bolts
9.7%
3.6×
Review
Adverse-weather imagery
6.5%
2.4×
Review
Substation thermal scans
4.3%
1.6×
Watch
Steel lattice structures
2.4%
0.9×
Normal
Bar: engineer-correction-rate lift vs. steel-lattice baseline · scale 0–4.0× · tick marks the 2.0× review threshold2 of 4 slices over threshold
Evidence-linked improvement
A miss closes as a test, not a note
A cycle is done when the miss has become a test the next release has to survive. That suite is what the next detection off the line is measured against.
Improvement cycle · five stagesSwitchback — the path turns at Improve and returns at Learn
01Detect
Engineer-correction rate rises in an asset slice.
02Diagnose
Not the model first — read the frames and the asset records behind the corrections until one criterion is left.
03Improve
Version-stamp the change and attach the detections that exposed it.
04Verify
The affected cases run again, and a fail stops the release.
05Learn
It becomes a standing test, and the inspection rules change with it.
Learn → DetectThe return edge. The next detection runs against a suite one case longer.
Typical build scope
Twelve workstreams across six weeks
The build scope read against the delivery timeline. Week structure follows the six-week plan — discovery, sources, detection workflow, evaluation, integration, then production validation and handover.
WorkstreamWeek 1Week 2Week 3Week 4Week 5Week 6
01Inspection workflow discovery and automation-boundary definition.
02Inspection and asset source assessment.
03Detection criteria, severity band and asset-class mapping.
04Imagery ingestion and normalisation.
05Detection logic and frame binding.
06Confidence scoring and coverage routing.
07Engineer disposition workflow.
08Asset- and work-management integration.
09Missed-defect regression cases.
10Guardrails and disposition controls.
11Asset-trail instrumentation.
12Deployment, documentation and Agent Care handover.
12 workstreams · 6 weeks · bar shows the weeks a workstream is active — several run in parallelFinal scope and sequence confirmed in discovery
Engagement tiers
What each tier includes
Rows are the capabilities named in each tier's scope. Higher tiers include everything below them.
Capability✓ in scope · — not at this tierPilotOne circuit, one asset classProductionProduction asset systemsAdvancedMultiple regions / programmes
Introduced at Pilot
Detection to your record and criteria✓✓✓
Engineer disposition✓✓✓
Detection-quality baseline✓✓✓
Introduced at Production
Reporting by asset class—✓✓
Disposition workflow in your systems—✓✓
Approved write-back—✓✓
Asset-management integration—✓✓
Introduced at Advanced
Multi-region criteria and severity rules——✓
Multi-stage engineering approvals——✓
High inspection volume——✓
Multi-region programme controls——✓
Build priceFrom $5,000From $8,000Custom quote
Final build priceConfirmed after discovery based on integrations, workflow complexity, transaction volume, approval controls and deployment requirements.
Separate from buildBuild pricing is separate from recurring Agent Care, which covers managed monitoring, evaluations, incidents and verified improvements after launch.
What we need from you
What you bring, and what we build with it
Each input maps to a piece of build scope and a week in the delivery timeline.
You bringWe build with it
01Your asset classes and inspection record structure→Imagery ingestion and asset mappingWeek 1
02Representative inspected assets→Detection baseline, criteria extraction and frame bindingWeek 2
03Your detection criteria and severity bands→Detection criteria, severity band and asset-class mappingWeek 1
04Access to relevant APIs, feeds or exports→Inspection and asset assessment, then integration setupWeek 2
05Detections you would not want acted on→Coverage cases and the evaluation suiteWeek 4
06What must reach an engineer before work is scheduled→Confidence scoring, coverage routing, guardrails and disposition controlsWeek 3
07Named engineers to disposition detections→Engineer disposition workflow, then pilot and production validationWeeks 5–6
Nothing else is requiredDeployment, documentation and Agent Care handover are ours.
Delivery timeline
Four phases across six weeks
Phases are drawn over the weeks they actually occupy, which is why week 5 doubles up rather than padding.
PhaseW1W2W3W4W5W6
DiscoveryW1
BuildW2 – W3
EvaluateW4 – W5
Pilot & LaunchW5 – W6
Week focusW1Inspection workflow discovery, criteria mapping and the automation boundaryW2Imagery and asset integration and the detection baselineW3Detection workflow, confidence logic and disposition controlsW4Evaluation suite, coverage checks and failure-mode testingW5Asset-system integration, pilot circuits and targeted correctionsW6One inspection cycle read under the asset team, then handover
Reading the bandEach bar sits on the weeks its own work runs, and no others. Week 5 genuinely carries two kinds of work.
At the end of W6The last checks clear on live inspections and monitoring moves to Agent Care.
DurationSix-week plan shown · typical delivery 4–6 weeks depending on scope confirmed in discovery.
Next step · Energy AI agent
Build an asset-maintenance agent that stops at the disposition.
Show us your inspection passes and the criteria your engineers already apply. You bring the named engineer who dispositions a finding and the programme it feeds; we set the automation boundary at that person.
