Does it act on its own?
Only where you have said it can. By default it watches, diagnoses and recommends, and a named person approves. When a step changes something in the real world, the plan is frozen with a fingerprint of its inputs before it goes for approval, so what runs is exactly what was reviewed.
Won't it bury us in alerts?
It is built not to. It stays quiet while a metric is inside its normal range, and it speaks up only when the drivers explain the move. A movement it can't explain is reported as unexplained, not dressed up as a finding.
Where do the metric trees and the business context come from?
The metrics come from the semantic layer you already govern, so there is one definition of margin for people and agents alike. Playbooks are packaged as skills. Facts are kept in project memory and carried from run to run. All of it lives in your own storage, versioned.
What does it cost to watch a metric around the clock?
Each check is a compiled workflow, so most of its steps run as code and call no model. On our benchmark workflow a run costs about a cent. You can see the cost per run while you set it up, and hold to it afterward.
Does it only look for problems?
No. The same trees that show a number slipping show one running ahead. It reports openings the same way it reports problems: what it sees, why, what it is worth, what to do, and who decides.