Run · Agent fleet
Every agent, one grid, six signals.
"How many agents do we actually have running, and is anyone looking after them" is a question that should have a one-screen answer, not a tour through a dozen separate configurations.
Why one grid
An agent nobody's watching is a liability, not an asset.
As a company adds more agents, the real risk isn't any single one misbehaving. It's one quietly going stale, unowned, uncorrected, while everyone assumes someone else is watching it. The fleet grid exists so that assumption is never necessary.
The six signals
Same six, on every agent, always comparable.
- Function
- What the agent actually does: its role, not just its name.
- Owner
- Who is accountable for it, with a sensible fallback if nobody was ever assigned.
- Status
- Active, paused, deploying, or archived. One glance, not a click into each agent.
- Outcome
- The health score from Agent outcomes, right on the same tile.
- ROI, 30 days
- What it has actually been used for, over the period that matters for a decision.
- Last activity
- When it last did anything. The fastest way to spot one nobody's using.
"Outcome" is explained fully on Agent outcomes.
FAQ
Companies growing past a handful of agents are asking
What are the six signals shown on every agent?
Function, Owner, Status, Outcome, ROI (return on investment) over the last 30 days, and Last activity. It is the same six for every agent, so any two are directly comparable at a glance.
Can we filter or sort the fleet by any of those signals?
Yes. Sort and filter by status, owner, function, or any of the other signals, both as a Company viewing its own agents and as a Partner viewing a client's.
Does pausing an agent in the fleet lose its configuration?
No. Pause stops it from responding. Everything about it, including guardrails, knowledge and history, stays exactly as configured until it is resumed.
See every agent your company runs, in one screen.
Twenty minutes with the founder. No slides. The real product.
