We ran our own charging network. This is what it taught us.

GridLink started as a fleet charging platform. We built and ran our own CSMS, which meant we also ran the chargers: the depots, the vehicles plugged into them, and every session that had to finish before the morning shift.

Then we hit the wall that everyone in charging hits. Chargers and vehicles that are supposed to interoperate fail each other in small, specific ways, and each pairing has its own edge cases. Troubleshooting one by hand teaches you nothing about the next one. There is no number of hand-coded fixes that catches up, and every one you miss turns into a site visit.

So we built a workflow engine to handle failures generically: read the OCPP stream, find the cause, run the repair, verify the charger came back. That engine became Otto, and it turned out to be the more valuable half of what we had built. We pivoted the company to it.