FOR SOLAR PLANT OWNERS AND O&M TEAMS

Less time processing inspections. More energy produced.

SkyOps turns every drone inspection into a prioritised list of what to fix. The AI does the processing, your team acts sooner and on the right modules, and the next flight proves each repair. You keep flying the drones you already use.

WHO FLIESYour team or your drone provider
WHO DECIDESThe AI suggests, your team confirms
WHAT YOU GETA prioritised list of what to fix

No drone team? We can arrange the flight with a certified partner.

3D DIGITAL TWIN · LIVE FLIGHTSolar Park A · Zone B1 · MIC-2026-04-033
Critical · hot cell+28 °C hotter than its neighbours · fire risk
TWIN SOURCESGeometry · DSM + mesh from this flightLayout · module registration 800/800State · report B1
HealthyMinorPlan a repairCriticalPID · power lossOffline stringSoiling
ALTITUDE 38 m AGLCOVERAGE 1%DETECTIONS 0SAMPLE DATA

THE PROBLEM

Every week a defect waits, it costs you energy.

Faulty equipment quietly eats into what every solar plant produces. The defects show up in the inspection images, but turning thousands of images into decisions is still mostly manual and takes weeks. Meanwhile the defects keep costing energy, some get worse, and nobody can say which one to fix first.

5.08%of a solar plant's output was lost to equipment faults on average in 2025
2×the losses of five years ago, and still rising for strings, combiners and trackers
≈ 86 MWhlost per MW a year by a typical plant in Portugal*

Source: Raptor Maps, Global Solar Report 2026, 373 GW of solar plants analysed. *Our estimate: 5.08% of 1,700 MWh per MW a year.

WITH SKYOPSSAMPLE DATA
P1
Offline string — check fusesWO-2026-0412 · Assigned · Electrician 01
19 MWh /yr
P1
Tracker stuck — rows 14–27WO-2026-0421 · In progress
42 MWh /yr
P2
Bypass diode cluster — Zone AWO-2026-0414 · In progress
6 MWh /yr
Energy lost to open work94 MWh /yr · 6 orders
  • Defects found and ranked by the AI, confirmed by your team
  • A report to the international standard, generated from the review
  • Repairs ranked by energy lost, assigned and verified by the next flight
WO-0377 · PID box on CMB 4.2 · verified by MIC-2026-04-03328 MWh/yr recovered

Energy values are estimates: what each defect stops producing, from the module, the sunlight and how severe the defect is.

HOW IT WORKS

You fly. SkyOps turns the images into decisions.

Three steps, and then the cycle starts again: every report becomes the starting point for the next flight.

  1. Area surveyWaypoints
    Z01Z02
    AREA180.4 ha
    DISTANCE85.9 km
    WAYPOINTS238
    PHOTOS4,243
    01 · YOU FLY

    Fly any drone, as you do today

    Your team or your drone provider flies, with the drones you already have, and uploads the thermal and colour images. SkyOps can prepare the flight plan from the site layout or from the last report's open findings.

  2. MODULES15,840
    ANOMALIES778
    ENERGY LOST≈ 690 MWh/yr
    ARGUS · AI12 repairs suggested
    02 · THE AI PROCESSES

    No more reviewing images one by one

    The AI processes every image for you: it finds each defect, matches it to its exact module, compares it with the last inspection, estimates what it costs and ranks what to fix first.

  3. Confirmed by your team58 detections
    Report shared with the ownerportal + PDF
    Repair done in the fieldElectrician 01
    Verified by the next flight28 MWh/yr recovered
    03 · YOUR TEAM DECIDES

    Your team acts on what matters first

    Your team confirms the priorities and sends the repairs to the field, starting with the defects that cost the most. The next flight confirms each repair, so nothing is left half done.

Sample data from a demo inspection.

INSPECTION REVIEW

Your team reviews decisions, not thousands of images.

Argus, the SkyOps AI agent, has already sorted every detection. Each one arrives with its thermal and colour evidence, its exact module, the reason it was flagged and a suggestion. Your engineers confirm or correct it and sign off. Nothing reaches a report without a human decision.

Portfolio / Solar Park A / MIC-2026-04-033 / 5 · ReviewSAMPLE DATA
Processing 68% · showing detections from zones processed so far
Inspection review58 detections · 64 pending · 0 accepted
Finalize · 64 pending excluded
ARGUS · AIArgus triage2 open
Accept 1 detection that matches confirmed signatures97%Why?DismissApprove
Reject 6 low-confidence, low-severity detections as noise81%Why?DismissApprove
Severity AllVs. last AllSource AllStatus PendingConfidence ≥ 0.60J/K navA acceptR rejectC reclassifyU assign
IDTYPESEVERITYVS. LASTCONFΔT
OBS-04200Hot cellMODULEcriticalNEW0.97+28.4 °C
OBS-04204Offline stringSTRINGcriticalNEW0.94+18.2 °C
OBS-04240Inverter/combiner faultSTRINGcriticalDEGRADED0.88+8.9 °C
OBS-04230DelaminationMODULEcriticalNEW0.91+8.4 °C
OBS-04263Bird droppingOBSTRUCTIONcriticalPERSISTENT0.62+25.1 °C
OBS-04215Tilt-tracker faultMOUNTINGcriticalRESOLVED0.85+13.0 °C
OBS-04201Activated bypass diodeMODULEhighPERSISTENT0.93+16.2 °C
OBS-04222Module crackingMODULEhighNEW0.79+9.0 °C
OBS-04221VegetationOBSTRUCTIONcriticalNEW0.90+29.0 °C
OBS-04214Racking issueMOUNTINGhighDEGRADED0.84+11.2 °C
OBS-04223Reverse polaritySTRINGhighNEW0.81+22.1 °C
OBS-04258Missing moduleMODULEhighNEW0.77+15.4 °C
12 of 64 pending · sorted by severityJ/K to move
OBS-04200Hot cellcriticalNEWINSPECTION_SYSTEM
ThermalRGBBoth
Thermal · H30T
RTK fix · ±2 cm
RGB · wide
RTK fix · ±2 cm
MODULEMOD-0119
STRINGINV-07.CMB-06.STR-09
ΔT28.44 °C
AREA1.29 m²
IRRADIANCE952 W/m²
ARGUS · AIWhy the model flagged thisACCEPT
SignatureSingle-cell hotspot · ΔT +28.4 °C · circularCorroborationSingle source · this inspectionConditions952 W/m² · wind 2.5 m/s · emissivity 0.85 · within IEC TS 62446-3Similar cases19 past detections · 99% confirmed by your team
LINEAGE
Ingest08:12Process · ai-v4.208:55Thermal segconf 0.91Asset matchMOD-0119Temporal diffnew
UnassignedAssign · UReclassify · CReject · RWork orderAccept · A
  1. 1

    Argus triage

    Bulk suggestions with a confidence score and a reason. Nothing changes until a reviewer approves.

  2. 2

    Review while it processes

    Start on the first zones while the rest of the plant is still processing.

  3. 3

    Filter what matters

    By severity, change since the last inspection, source, status and model confidence.

  4. 4

    Evidence side by side

    Thermal and RGB of the same module, georeferenced to its exact coordinates.

  5. 5

    Why the model flagged it

    Thermal signature, corroboration, capture conditions against IEC TS 62446-3 and how often similar cases were confirmed.

  6. 6

    Full lineage

    Every detection traces back through ingest, model version, segmentation, asset match and temporal diff.

  7. 7

    One step to action

    Accept, reject, reclassify, assign or open a work order, all from the keyboard.

WORK ORDERS · CLOSE THE LOOP

Fix what costs the most first, and prove it worked.

Confirmed defects become work orders your crews and contractors can act on, with the energy each one is costing you on every card, so the biggest losses get fixed first.

Prioritised by energy lost

The AI proposes work to create or re-prioritise, each with the energy at stake.

Assigned to the right people

Your O&M team or contractors, with due dates and a field checklist.

Verified from the air

A fix only closes when the next flight confirms the defect is gone.

Energy recovered, measured

See how much energy completed and verified work brings back every year.

Operate / Work ordersSAMPLE DATA
OPEN694 MWh/yr at risk
OVERDUE1past due date
AWAITING VERIFICATION1closed by next inspection
ENERGY RECOVERED32 MWh/yrdone + verified
ARGUS · AIReplace the 7 hot-cell modules left out of WO-1312.4 MWh/yrApprove
Open2
P1WO-0413
Hot cell ΔT +28 °C — replace moduleSolar Park A · Zone A
—due 04-230.3 MWh/yr
P3WO-0415
Robotic wash — soiling sector CSolar Park A · Sector C
C1due 05-1525 MWh/yr
Assigned2
P1WO-0412
Offline string — check fuses & re-energizeSolar Park A · INV-08
E1due 04-1919 MWh/yr
P1WO-0420
Hot cell cluster — inverter 3Solar Park B · INV-03
T1overdue2.1 MWh/yr
In progress2
P2WO-0414
Bypass diode cluster — Zone ASolar Park A · Zone A
T1due 04-306 MWh/yr
P1WO-0421
Tracker stuck — rows 14–27Solar Park C · Rows 14–27
T2due 04-2042 MWh/yr
Done1
P3WO-0398
Vegetation clearance — NW tree lineSolar Park A · Sector N
T2verify next flight4 MWh/yr
Verified1
P2WO-0377
PID — install anti-PID box on CMB 4.2Solar Park A · INV-04
E1verified 04-1528 MWh/yr
WO-2026-0412P1
Offline string — check fuses & re-energizeSolar Park A · INV-08 · CMB-04 · STR-11
  1. DetectedOBS-04204
  2. WO created2026-04-16
  3. AssignedElectrician 01
  4. 4Field fixpending
  5. 5Verifiedby MIC-2026-07-041
LOSING19 MWh/yr
ALREADY LOST0.2 MWh
DUEin 3 days
CHECKLISTIsolate combinerTest string fusesRe-energize & IV trace
Start work

WHY SKYOPS

Other tools stop at the report. SkyOps stops at the repair.

There are good inspection tools and good drone service companies. Most of them hand over findings and move on. SkyOps is built for what happens next: deciding what to fix first, getting it fixed and proving it worked.

Where it ends

TYPICAL TOOLS
At the map and the PDF report.

SKYOPS
At a repair confirmed by the next flight.

Which drones

TYPICAL TOOLS
Often tied to one drone brand or one inspection service.

SKYOPS
Any drone, flown by your team or the provider you already use.

How the AI decides

TYPICAL TOOLS
Detections without an explanation.

SKYOPS
Every suggestion comes with its reason. Your team makes the call.

What you are losing

TYPICAL TOOLS
A list of defects.

SKYOPS
Each defect with the energy it is costing you, so the biggest losses get fixed first.

Remote sites

TYPICAL TOOLS
Needs a good connection in the field.

SKYOPS
Works offline in the field for seven days or more.

Time to act

TYPICAL TOOLS
Weeks of reviewing images by hand before anyone acts.

SKYOPS
The AI does the processing. Your team only reviews the decisions.

Built for the people responsible for the plant's output.

PLANT OWNERS · ASSET MANAGERS

Know how much energy each defect is costing, act on the right ones first and see what every repair recovered, across every plant. Written for decision makers, not only for engineers.

O&M COMPANIES

Less time processing inspections, more time fixing. Your crews get a prioritised list of work orders, and every fix is proven to the owner from the air.

Drone inspection company? SkyOps also helps you deliver faster reports and repair tracking to your clients, with your pilots, certifications and drones in one place.

FAQ

Straight answers.

Does SkyOps fly the drones?

No. SkyOps is software. Your team or your current drone provider flies, with the drones you already have, and uploads the images. If you have no one to fly yet, we can arrange the flight with a certified partner.

What is Argus?

Argus is the AI agent built into SkyOps. It goes through every inspection image, finds and prices the defects, and suggests what to do about each one, with the reason. It never approves anything on its own: your team makes every decision.

How do you estimate the energy lost?

Each defect is turned into the energy it stops producing, based on the module, the sunlight at the time of the flight and how severe the defect is. We measure in energy, not euros, because energy prices vary from plant to plant. Values on this page are sample data.

Who is SkyOps for?

Plant owners and O&M companies: the people responsible for keeping a plant producing. Drone inspection companies can also use it to deliver reports and repair tracking to their clients.

How do we get started?

With a demo. Then one plant, one inspection cycle: you upload a recent flight, or we arrange one. SkyOps delivers the report and a priced repair list, and the next flight checks which repairs were done.

See SkyOps on one of your plants.

Book a demo and bring a recent flight if you have one. We will show you the prioritised list of what to fix it produces, and how the next flight proves what was fixed.

* Required. Sending opens your email app with the request ready to go. Or write to contact@marcdes.com