Agent fleet
The agents running in the background, what they did, and what they are waiting on.
Categorisation, reconciliation matching, document extraction, collections chasing, anomaly detection. Each runs on its own cadence and each produces a record.
An agent that has stopped is more dangerous than one that never ran, because the work silently stops being done.
Where everything sits
How to work this page
Last run, next run, and the outcome. A stalled agent is the failure mode to watch for.
Actions taken, items flagged, and anything handed to a human.
Anything the agent could not settle goes to Nerve. A growing queue means the thresholds need adjusting.
Where an agent acts automatically versus asks. Too cautious creates work, too confident creates errors.
On a phone
Every figure from the desktop appears here, stacked rather than reduced. Tables scroll inside themselves so the page never moves sideways, and figures keep their separators and their alignment at every width.
Questions people actually ask
No. Agents prepare and flag. Payment execution requires human approval without exception.
It reports the failure with its cause and stops rather than continuing on bad state. Nothing downstream runs on its output.