Argus started as an internal tool. A field engineer was driving between wells to read numbers off a drive display, and losing a shift of production every time a pump tripped at 2 a.m. and nobody knew until the morning round. The first version existed to stop that happening on his own wells. It still runs on them.
The producing wells in the case study are operated by Green Flame Resources, and Argus was built inside that operation by the engineer who runs them. That is not an arm's-length customer reference and we do not present it as one.
What it does mean is that the software was shaped by consequences rather than by a specification. Every alarm exists because something happened at 2 a.m. and the alarm did not. The torque-creep detection exists because a well sanded up slowly enough that fixed thresholds never fired. The offline queue exists because Starlink blinks.
The next fields will be somebody else's, and we will label those differently when they arrive.
First well live in days, not quarters.