The terms we use, in plain words
Governance and engineering use the same words differently. These are the meanings we use across the platform and our research.
- Accountable executive
- The named person answerable for an AI system, usually the holder of the related FAR key function or business accountability.
- Action gate
- The point where Cogna8 decides whether an agent may take a proposed action: allow, needs approval or block.
- AI system
- Any model, agent, endpoint, copilot or MCP server an organisation runs or relies on, including AI inside vendor software.
- Agent
- An AI system that can take actions, such as sending messages, calling tools or changing records, not only produce text.
- Conflict
- Two sources disagreeing about a fact an agent relies on, for example two different claim amounts. Cogna8 records conflicts instead of overwriting one value.
- Control
- A specific measure that satisfies an obligation, with an owner, the AI systems it governs and an implementation level.
- Decision receipt
- The record of one gate decision: what was asked, what was known, which rule applied and the outcome. Kept as evidence.
- Deterministic decision
- A decision that does not depend on a model judgement, so the same inputs always produce the same outcome and it can be replayed.
- Governed state
- Facts an agent relies on, held as typed, versioned items with their source, instead of loose text in a prompt.
- Impact assessment
- A recorded assessment of how an AI system could affect people, customers or operations, kept with the system it covers.
- MCP server
- A Model Context Protocol server: a standard way for agents to reach tools and data. Cogna8 tracks which agents can reach which MCP servers.
- Obligation
- A requirement from a framework or law, with its source and effective date, that one or more controls satisfy.
- Proof levels
- How far a control is proven: declared, observed, implemented, evidenced, tested and effective. Effective is never inferred without a passed test.
- Risk tier
- The level set by a written rule over an AI system's risk dimensions. It decides which controls apply and how often the system is reviewed.
Let's connect
Tell us where your AI programme is today. We start with a short, scoped pilot on your own AI estate, and every enquiry is handled in confidence.