The method, end to end
Good inspection follows the same path every time, and the discipline is the point — it is what turns a wall of data into a short, ranked list of things worth doing.
| Step | What happens | Why it matters |
|---|---|---|
| 1. Connect every source | Read every major inverter brand’s cloud data — no new hardware | Nothing hides between separate apps |
| 2. Score against expectation | Compare each string to a weather-normalised expectation, daily | A fair yardstick, not the nameplate or last year |
| 3. Classify by severity | Grade findings informational / warning / critical (IEC 61724-1) | The signal is separated from the noise |
| 4. Verify (AI + engineers) | AI grades at scale; solar engineers check the critical ones | A real fault reaches you; a passing cloud doesn’t |
| 5. Attribute & act | Name the likely cause and rank by size | The biggest recoverable issue is fixed first |
No new hardware, every brand
Inspection connects to the data your inverters already produce in their manufacturer clouds, so there is nothing new to fit on site. Reading every major brand into one independent view is what removes the blind spots — the fault that would otherwise sit unnoticed in the one app nobody opened this week.
AI and engineers, together
Scale and judgment both matter. AI grades every string every day — far more than any team could review by hand — and solar engineers verify what the grading flags as critical before it becomes an alert you act on. This is the “technology and people” pairing: the machine finds the candidates, the engineer confirms the call. It is also our answer to a market full of faceless “AI platforms”.
Why the method makes the output defensible
Because the steps are standard-aligned, independent, and human-verified, the result holds up where it counts: a warranty claim, a dispute with an EPC, or a buyer’s due diligence. The value isn’t only that a problem was found — it’s that it was found by a method someone else can trust.
