Platform · Governance & Compliance
Built to be governed, not merely monitored.
Governance & Compliance is the posture the whole platform enforces in code: default-deny access, an append-only audit log, maturity gates, spend caps, and kill switches.
Executive overview
These controls are not a product tier — they are how the platform behaves for everyone. Access is default-deny and checked server-side on every request. Every privileged action writes an append-only audit event with actor, organization, target, and timestamp, and nothing in the application deletes or rewrites one.
Enforcement is structural rather than procedural. Spend and usage caps are hard limits checked at approval time, not alerts reviewed after the fact. Maturity gates make unready products unlicensable in code. And any AI role can be paused or revoked instantly, with no cascading cleanup.
Core capabilities
Default-deny access
Capabilities are granted explicitly; everything ungranted is refused server-side.
Append-only audit
Actor, organization, target, timestamp — written on every privileged action, never rewritten.
Maturity gates
Products below licensable maturity cannot be sold or deployed to production.
Hard caps
Usage and spend limits enforced at approval time, not reported at month end.
Kill switches
Any role, any deployment: paused or revoked in one action.
One honest boundary: the controls above run in production today, but the dedicated compliance-review console — legal holds and review workflows on a single screen — is specified and in design. This page shows no screenshots of it because it does not exist yet; when it ships, what it surfaces will be the enforcement that already runs.
Why enterprises care
Evidence by default
An append-only log means compliance evidence accumulates as a side effect of operating, not as a quarterly project.
Bounded exposure
Hard caps convert “what could this cost us?” from an estimate into a configured number.
A reversible decision
Kill switches make adopting an AI role something you can undo in one action — which is what makes it approvable in the first place.
How it governs the workforce
This module is why “AI employee” is a governed claim rather than a metaphor. The four commitments every workforce page cites — permissioned, audited, capped, stoppable — resolve to these controls, and the public record of that promise lives on the trust page.
- Append-only audit
- Hard spend caps
- Instant revocation
Across the ecosystem
See it run a workforce.
A demo walks the module on live platform data — no slide deck.