The terms, in plain English.

Agent Passport®

The way an organisation decides who is allowed to do what with AI. Each person holds one of three levels, based on what they have shown they can do. It is an internal control classification, not a certificate or a badge. More on the Agent Passport®

AI agent

A piece of AI software that takes actions, not only makes suggestions. For example, an agent might read a shared mailbox and draft or send replies. Because it acts, what it is allowed to reach matters as much as what it says.

AI governance

The working arrangements that decide which uses of AI are permitted, who approves them and what records are kept. A policy says what people should do. Governance decides what they are permitted to do, and produces the evidence.

AI Impact Assessment

A structured set of questions asked about a proposed use of AI before it goes ahead: what it does, what data it touches, and what could go wrong.

AI literacy

Staff having enough knowledge and skill to use AI sensibly in their role. Article 4 of the EU AI Act requires organisations that use AI to take measures to support it. It has applied since 2 February 2025.

AI register

One current list of every AI tool, agent and use in the organisation, with a named owner beside each entry.

Business Owner

The named person accountable for a use of AI once it is live: its value, how it runs, and when it is retired.

Control gate

The point where a use of AI is checked before it is rolled out more widely. Three checks: is this person passported, is there evidence, and has someone accountable approved it?

Control system

The access rules, evidence requirements, approval workflow, register and reporting that let an organisation adopt AI safely. It is what Ambient Agent® installs.

Copilot

Microsoft Copilot, formerly Microsoft 365 Copilot, is Microsoft’s AI assistant, built into tools such as Outlook, Teams and Word. Staff can also use it to build their own agents, and in many organisations they do not need a paid Copilot licence to do so.

Deployer

The term the EU AI Act uses for an organisation that uses an AI system in its work, as opposed to the organisation that made it.

EU AI Act

Regulation (EU) 2024/1689, the European Union’s law on AI. It applies by where an AI system is placed on the market or used, so it can reach UK organisations. See the dates

Governance Approver

The named person who approves risk, policy and compliance decisions at the control gates.

High-risk AI system

A category in the EU AI Act for AI used in sensitive areas. These systems carry the heaviest duties: risk management, documentation, human oversight and logging.

ISO/IEC 42001

An international standard that sets out how an AI management system should be structured and run. An organisation can be certified against it.

NIST AI RMF

The AI Risk Management Framework published by the US National Institute of Standards and Technology. A widely used voluntary guide to managing AI risk.

Provider

The term the EU AI Act uses for an organisation that develops an AI system or places it on the market.

Sector overlay

An extra layer added to the control system for your sector. It maps what your regulator expects to the part of the system that shows it. See the sectors

Shadow AI

Any AI use inside an organisation that nobody formally approved, owns or records. It is rarely deliberate. It is the ordinary result of capable tools arriving inside licences people already hold.

Six engines

The six working parts of the control system: Diagnostic, Capability, Control, Value, Reporting and Communications. Control sits at the centre.

No term matches that. Try a different word.