Register your first AI system
The AI system is the anchor of everything in Gamut. It is registered in AI System Records, and registering one well makes every later step, intake, risk tiering, assessment, evidence and reporting, faster and more consistent. Nothing else can attach to a system until the record exists: the registry is what intake and assessment feed from.
Work inside a workspace
Section titled “Work inside a workspace”Everything you create belongs to a workspace (also called an assessment), the container that scopes your systems, intake, evidence and reports. Make sure you have created or opened a workspace before you begin; AI System Records lists the systems within the workspace you are in.
What counts as an AI system
Section titled “What counts as an AI system”Register anything that uses AI and carries governance relevance:
- Internal applications and services that embed a model.
- Vendor or third-party AI tools your teams use.
- GenAI assistants and copilots in active use.
- Agentic workflows that take action through tools and APIs.
If in doubt, register it. An under-governed AI system is a bigger risk than an extra record. If a system was surfaced by the AI Discovery Inbox, you can promote it into AI System Records rather than entering it from scratch.
Fields to capture
Section titled “Fields to capture”A system record is structured. You do not need every field on day one, but the more you capture, the better intake and routing work later:
| Group | Fields |
|---|---|
| Identity | Name, system ID, version, lifecycle stage (development through retired). |
| Ownership | Owner and owner email, department, responsible-AI contact. |
| Nature | AI technique, model type, autonomy level, human-oversight model, vendor, deployment type and environment. |
| Data & exposure | Data classification, personal-data involvement, data sources, users, affected persons, geographies, regulatory exposure. |
| Governance | Risk tier, ACRS score and review dates (these are set or confirmed later, in intake and risk tiering). |
- Open AI System Records in your workspace.
- Add a new system record (or promote one from the AI Discovery Inbox).
- Enter the identity and ownership fields: name, owner, owner email, department.
- Describe the nature of the system: what technique and model type it uses, its autonomy level and how humans stay in oversight.
- Record the data and exposure context: data classification, personal data, who is affected, and where it operates.
- Save. The system now appears in your inventory and is ready for intake.
Good practice
Section titled “Good practice”- One owner, always. Ownership is what makes governance actionable.
- Describe purpose, not implementation. Reviewers care about what the system does and why.
- Register early. Capturing a system at proposal stage is far cheaper than reconstructing its history later.
- Let discovery feed the registry. A registry that misses systems produces governance that only looks complete.
With the system registered, complete AI Use Case Intake and set a risk tier, then run your first assessment.