AI Ready Pathway

Develop an AI System Inventory

Desktop computer showing a window with box icons, and a magnifying lens inspecting a package box.

Create and regularly update a central list of AI systems used or developed in your organisation. You may already have an IT asset register or catalogue you can leverage for this purpose.

This inventory should include basic details such as the purpose of the system, the owning team, risk level, deployment status, and whether the system uses machine learning or rules-based automation.

For organisations where environmental factors are material to risk assessment, the inventory should capture relevant environmental metadata, such as the computing location (e.g. data centre, edge device).

The inventory forms the backbone of governance by enabling visibility, accountability, and prioritisation of oversight efforts.

Two question marks inside speech bubbles, overlapping.

Why it matters

An accurate inventory ensures that no AI use goes unnoticed or unmanaged. It enables proportionate risk management, supports compliance reporting, and lays the groundwork for applying tiered governance controls. Without an inventory, organisations are effectively blind to their AI exposure. An accurate inventory gives teams a clear view of what systems are in use, supports risk tracking, and lays the foundation for proportionate governance.

Text "quick tips" inside a lightbulb icon, surrounded by funky, sparkly shapes.

Implementation tips

  1. Start with a simple spreadsheet or use an existing asset registry tool.
  2. Define required metadata fields and keep them lightweight for ease of entry.
  3. Classify systems by type, function, business unit, and risk level.
  4. Include both internal builds and third-party AI services, and tag them as such.
  5. Make sure the inventory is visible to relevant teams (e.g. privacy, security, legal, sustainability).
  6. Link inventory entries to risk assessment where possible.
  7. Assign responsibility for keeping the inventory up to date.
  8. When environmental factors are material, tag systems that may require environmental impact assessment (e.g., training large models, high-volume inference systems).
One hand passing a heart to another.

Support materials

Institute of Community Directors – AI Governance Framework Template
A starter template with guidance on documenting and overseeing AI systems.

New York State ITS – AI System Inventory Guidance
Practical guidance on how to collect and structure AI system metadata.

Ethos – Building Your AI System Inventory
Simple guide to starting and evolving an AI system inventory.

ZealStrat – AI System Inventories: The Foundation for Governance
Explains the role of system inventories in scalable AI governance.

printer icon