60%of switching errors caught before the field
The problem
A large utility writes tens of thousands of switching orders a year. Each is checked by a second person reading it line by line, and the checks that matter most (load at every intermediate step, a missing closure, the wrong device or tag) are the hardest to see on paper. When a switch fails mid-procedure, there is rarely a written fallback.
What we ran
A large distribution network model, a year of routine and complex orders, and planted failure modes taken from operators: a missing closure, the wrong device or tag, an intermediate overload, cut jumpers and loop normal-opens.
What the agent did
It drafts routine orders from topology and past orders, runs load flow at every intermediate step against the utility's limits, and writes a fallback for each switch that could fail to operate. The dispatcher owns every order. Complex orders go to a person with the reasons.
Results
| Measure | Today | With Evardi |
|---|---|---|
| Errors that get past the second check | 5 a year reach the field | 2 a year (60% caught) |
| Minutes to review a routine order | 12 | 6 |
| Intermediate steps load-checked | Spot checks | Every step |
What it is worth
- Utility cash: orders a year × minutes saved × loaded dispatcher hour × share realized
- Staff capacity: the same hours, counted as dispatcher positions
- Customer value: errors caught × cost of each event, on its own line
* Study simulated with partners on public data.
Run it on your network
Last year's switching orders as a CSV, plus the feeder model. We return every order re-checked step by step, with the problems it found.
Book a demoSee Switching on the platform
Full study
Request the full study.
The complete write-up of “Every intermediate switching step checked, review time halved”: method, data, results and assumptions, as a PDF. Leave your work email and we'll send it.
