Observe
An estimate before deployment, not a surprise afterwards.
Before an agent is deployed the platform estimates its monthly cost from its trigger, its model and the work it is likely to do. After deployment you see the actual against that estimate. If the two diverge badly, that is a bug in the estimate and we would rather you saw it.
Shown at the review screen
Runs per month, model usage, runtime, and a total. On the same screen as the permissions, because both are things you are agreeing to.
One figure, not a token model
You should not have to estimate token consumption to predict a bill. The platform does that arithmetic and takes the risk of having done it wrong.
Actual against estimate, per agent
Each agent shows what it was predicted to cost and what it did cost. An agent drifting above its estimate is usually an agent whose job has quietly grown.
A cost ceiling is a budget, and budgets stop things
Cost limits are enforced by pausing the agent, not by a warning email after the fact. See budgets and limits under the runtime.
Observability
The rest of this area.
Run timelines, reasons, metrics and cost — in business terms.
The run timeline
One run, step by step, in the order it happened.
Why did it do that?
Any action, expanded into the records and rule behind it.
Exception-based management
Agents handle the routine. You handle the exceptions.
Metrics
Tasks, approvals, outcomes. Model usage where it belongs.
Impact
Thirty days of one agent, in figures you can check.
Give AI a job.
Build an agent, choose what it can access, define what it may do, and deploy it to work safely across your business.
No code required · Full SDKs and API · Stop any agent instantly · [email protected]