An AI field engineer for satellite fleets. It finds the cause before anyone gets on a boat.
A ship drops for five minutes every two hours.
No alarm fires on it.
Eight more ships, on all four beams, do the same.
They all run the same firmware.
Each ticket arrives with a diagnosis, a severity and one next action.
The rest stay sealed until launch.
Request access →FieldAgent on our hardest benchmark. Right cause and right action.
Hidden second faults it also caught on the same tickets. We are working on the other three.
Changes it can make to a device. It recommends. Your engineers decide.
The benchmark is a simulated fleet of 40 GEO VSAT terminals. The next test is closed tickets from real operators.
The same reading, run across the whole fleet on a schedule. It finds the faults no one has reported yet.
Release date
The questions every NOC asks first, answered before you ask.
It reads from your NMS, your hub logs and your terminal status. It uses an API where there is one, and exports or a log store where there isn't. Nothing is installed on the terminals. Your hub platform decides which connector comes first, and we build it with you during the design partnership.
We host it to start. It can run in your cloud account or on your own servers when your data rules require it. We agree on this with each design partner before any data moves.
Your data is used only to diagnose your tickets, and it isn't used to train models. For the replay you send anonymized closed tickets, and ship names and customer details can be removed before anything leaves your network.
It can't change a terminal or a config. In shadow mode it writes its diagnosis next to your engineers' and acts on nothing. After that, every action is a recommendation that an engineer approves.
For satcom operators and service providers with a NOC that supports maritime or aviation terminals.