Home / What we do / Agent Passport®
Who can do what with AI.
The Agent Passport® sets each person’s AI access by what they have shown they can do. Everyone starts with governed use of approved tools. Access widens as capability is demonstrated.
| Level | What it allows |
|---|---|
| 1 Consumer | Uses approved AI and the agents shared with them. Cannot build. |
| 2 Practitioner | Builds simple agents, with approval. |
| 3 Builder | Builds advanced agents, with full sign-off. |
Three levels, each one earned.
Each level opens a wider set of AI tools: using AI, then building simple agents, then building advanced ones. At every level, an agent still needs evidence and approval before it is shared. A person’s level is assessed before any permission changes.
Consumer
Uses approved AI assistants and the agents shared with them, within set limits. Cannot build.
Where everyone startsPractitioner
Builds simple agents with the no-code builder in your approved tools, after the training for that tool.
Earned through trainingBuilder
Builds advanced agents, including ones that connect to other systems or act on their own, after the training for those tools and a record of approved agents at Level 2.
The highest levelIn a Microsoft estate, that is Copilot and shared agents at Level 1, Agent Builder at Level 2 and Copilot Studio at Level 3. The same pattern applies on other platforms.
An internal control classification. Your organisation assesses its own people. The training for each tool is the provider’s own. Ambient Agent® does not provide it.
The name invites one misunderstanding.
So it is worth being exact about what the Agent Passport® is.
It is
- An internal control classification.
- Assessed by your own organisation, on your own people.
- Installed by Ambient Agent® as part of the control system.
It is not
- A certification or a qualification.
- A credential or a badge.
- A register of individuals.
Two accountable roles, appointed by name.
Levels recognise capability. Accountability is separate, and it is appointed. Two named roles hold the decisions that keep the system defensible.
Business Owner
Accountable for a use of AI, or an AI agent, once it is live: its value, how it runs, and when it is retired. Every live use has a named owner.
Governance Approver
Approves risk, policy and compliance decisions at the control gates, and sits on the group that oversees AI in your organisation. This is the named answer to the question regulators now ask: who approved this, and on what evidence?
Why access follows capability.
- Access is mapped to demonstrated capability. Seniority and enthusiasm do not count.
- Every permission change is assessed against a person’s level before it is approved.
- Accountability is personal. Every AI user holds a level with defined limits.
- Your team reassesses every six months, so progress is tracked.
Where it starts.
AI Control Diagnostic. A first view of your workforce’s AI capability, using the Agent Passport® framework. It shows the priority gaps.
Control System Install. The full assessment of what your people can do with AI. It establishes who can use AI, at what level and to what effect, before anything is rolled out more widely.
Every six months. Your team reassesses.
Is AI access in your organisation open by default?
One of the ten Self-Assessment questions asks exactly that. It is a good place to start.