PromenautWhat an AI agent governance platform must do
A fair checklist for evaluating any platform:
Authority
An agent acts on the authority of the person who launched it, within their permissions, not as its own privileged account.
Boundaries
What an agent may read, write and call is declared and enforced while it runs, with a core of controls that stays on.
Human control
Sensitive steps wait for a person, maker is kept apart from checker, and work can be stopped cleanly.
Evidence
Each run leaves a tamper-evident record of who acted, on whose authority, against what they were approved to do, and which model answered.
Cost
A run’s cost is known before it starts and held to a budget, not discovered afterwards.
Data residency
The platform runs where your data already lives, so sensitive data stays in your estate.
Two ways to govern agents: an overlay, or by construction
A governance overlay
Observes, monitors and applies policy to agents built and run elsewhere. Valuable for visibility and policy across a mixed estate, and several good products do this well. The governance sits around the agents, so it depends on what it can see and intercept.
Governance by construction
The platform that runs the workforce also governs it. Authority, approvals, evidence and cost come from the work itself, because they are part of how the work is done rather than a layer around it.
An overlay can tell you what agents did. Governance by construction decides what they may do, and shows it, as the work happens.
Promenaut: a governance platform that is also the workforce
Promenaut governs AI agents by construction, because it is the platform that runs them: module agents reach the platform through one governed seam, enforced in the build.
In the same run, an agent reaches knowledge as a deputy of the person who launched it; sensitive steps wait for a person, and a sensitive knowledge change waits for a checker who did not launch it; the run is costed before it starts; each agent run records which model answered; each workflow run ends with a verdict in a tamper-evident ledger; and an operator can pause an agent or stop a running workflow.
It can run in your own cloud account from a shipped kit, including an environment with no outbound internet access, on the model providers you choose.
Promenaut is running in production with leading industry partners, including a global tier 1 bank.
Common questions
What is an AI agent governance platform?
The layer that controls what AI agents may do, checks they stay within it while they run, and shows what they did, so an organisation can run agents in production, especially under regulation.
What should an AI agent governance platform do?
Act on the authority of the person who launched the work, enforce declared boundaries, keep people in control of sensitive steps with maker and checker apart, leave tamper-evident evidence, know a run’s cost before it starts, and run where your data lives.
What is the difference between an overlay and governance by construction?
An overlay observes and applies policy to agents built elsewhere, so it governs from around the work. Governance by construction comes from the platform that runs the work, so controls and evidence are produced by the run itself.
Do banks need an AI agent governance platform?
To put agents into production, yes: a regulated firm has to show who acted, on whose authority, what was done, what it cost, and that the work can be stopped.
