The governed-AI glossary.
Governed AI has its own vocabulary — runtimes that gate every action, identities for agents, ledgers that prove what happened. This glossary defines the terms a regulated-industry buyer needs, in plain language, with how each one maps to a real Cortex capability. No jargon left undefined.
28 terms · governance · agents · provenance · operations · security
How agent actions are allowed, held, or refused.
The decision layer of the runtime — the gates, policies, and oversight that turn a written rule into an enforced verdict.
Governed runtime
The execution layer where every agent action passes a policy gate before it can run.
Governance gate
A single decision point that allows, holds, or blocks a proposed agent action.
Fail-closed
When a control can't confirm an action is safe, it denies rather than allows.
Policy-as-Code
Governance rules written as testable, versioned code that the runtime enforces directly.
Oversight modes
Per-agent autonomy levels that decide whether an action executes now or waits for a human.
Approval gate
A hold that pauses a high-risk action until a designated human approves it.
Break-glass
An audited override that lets an authorized human force a held action through, with a recorded reason.
Risk tier
A classification of how consequential an agent or action is, used to set the controls it needs.
Treating agents as governed enterprise identities.
Who an agent is, what it may do, and who is accountable for it — least privilege applied to autonomous software.
Agent IAM
Identity and access management that treats each AI agent as a governed enterprise identity.
RBAC
Role-based access control — permissions granted by role rather than to individuals.
Tenant isolation
Strict separation so one customer's data and agents can never reach another's.
Agent Studio
Where governed agents are built, chat-tested, and published — behind a reliability gate.
Where facts come from, and how records stay honest.
The data model agents reason over and the tamper-evident record that proves every fact traces to a real source.
Provenance
The recorded origin of a fact or output — what source it came from, and where.
Trust Ledger
A tamper-evident, hash-chained record of every governed run, action, and approval.
Hash chain
A sequence of records where each one cryptographically includes the hash of the previous.
Tamper-evident
A record whose integrity can be proven — any change to it is detectable after the fact.
Ontology
The business object model — types, instances, and relationships — agents reason over.
DLP
Data loss prevention — screening data in and out of tool calls to catch leakage and injection.
Running, measuring, and stopping a fleet of agents.
The day-two surface: observing quality, gating releases, capping cost, and containing an agent that goes wrong.
Action Fabric
A lifecycle that makes every agent action a governed object: dry-run, propose, approve, execute, compensate.
Control Tower
The live operations console for the whole agent fleet — watch, gate, pause, or kill.
Kill switch
An instant control that stops an agent or tool server from acting, immediately.
Cost governance / hard cap
Spend limits on agents and tenants, with a hard cap that blocks a run before it overspends.
Reliability score
A 0–100 quality score per agent that gates which versions are allowed to publish.
Eval suite
A set of graded test cases that score an agent's quality and gate releases against regressions.
Shipping governed solutions and the tools agents call.
How a governed solution moves between environments, and how the external tools an agent uses are themselves governed.
The evidence that makes governance provable.
The signed, verifiable artifacts that let you prove — not just claim — that the controls fired.
These aren't buzzwords — they're enforced controls.
Every term in this glossary names something Cortex actually does at runtime. A governance gate is a real verdict (402 / 403 / 409 / 451), the Trust Ledger is a hash-chained record you can verify offline, and provenance pins each fact to its source. The vocabulary is precise because the enforcement is.
- Each definition links to the capability that implements it
- Frameworks are framed honestly — aligned with, never certified
- Evidence is generated from real runs, sealed in the Trust Ledger
See the words become controls.
Book a walkthrough and watch a governance gate return a real verdict, the Trust Ledger verify offline, and a fact trace back to its source — the glossary, running live.