AI System Inventory: [Organisation]
Operational register of every AI system the organisation develops,
buys, embeds or uses, with role, legal classification, internal risk
tier, accountable owner, lifecycle status and review dates. Structure
aligned with the inventory expectation in GOVERN 1.6 of the NIST AI
Risk Management Framework (AI RMF 1.0, January 2023) and with the
scope, resources and documented information clauses of ISO/IEC
42001:2023 (structure only, original wording), and records roles and
risk categories used by the EU AI Act (Regulation (EU) 2024/1689 as
amended by Regulation (EU) 2026/1744). Replace every [bracketed]
value and every row marked Example. This is a documentation format,
not legal advice: check the current official texts and a qualified
professional.
Document control
| Field | Value |
|---|
| Document ID | [AIINV-001] |
| Organisation and scope | [Legal entities, regions, business units covered] |
| Version | [0.1] |
| Status | [Draft / Live / Under review] |
| Register owner | [@ai-governance-lead] |
| Contributors | [@procurement], [@it-asset-management], [@legal], [@dpo], [@security], [@product-leads] |
| Last full review | YYYY-MM-DD |
| Next full review | YYYY-MM-DD |
| Linked registers | [Applicability register ID], [Vendor register], [Records of processing] |
| Classification | [Internal / Confidential] |
Approvals
| Role | Name | Decision | Date |
|---|
| AI governance lead | [@name] | [Approve] | YYYY-MM-DD |
| Chief risk officer | [@name] | [Approve] | YYYY-MM-DD |
Revision history
| Version | Date | Author | Change summary |
|---|
| 0.1 | YYYY-MM-DD | [@name] | First discovery round |
| 1.0 | YYYY-MM-DD | [@name] | Example: Seven systems registered and classified, approved by risk committee |
| 1.1 | YYYY-MM-DD | [@name] | Example: AIS-007 retired; AIS-004 version updated |
Contents
1. Purpose and scope
Definition used: [Example: the AI Act Article 3(1) definition, a
machine-based system that operates with some autonomy and infers from
input how to generate outputs such as predictions, content,
recommendations or decisions, read with the Commission guidelines on the
AI system definition. Simple rule-based scripts with no inference are
out of scope but listed in the borderline log.]
| In scope | Out of scope |
|---|
| Example: Systems built in-house, bought, licensed as SaaS, embedded in other software, used through an API, used by staff on their own initiative for work | Example: Rule-only automation with no inference, personal non-work use |
Why the register exists: [Example: to know which systems trigger legal
duties, to assign accountable owners, to plan reviews, and to answer
customers, auditors and regulators quickly.]
2. Discovery sources
| Source | Method | Last run | Owner |
|---|
| Procurement and vendor contracts | Example: Keyword search for AI, machine learning, model, generative | YYYY-MM-DD | [@procurement] |
| SaaS and IT asset inventory | Example: Review of feature release notes for embedded AI | YYYY-MM-DD | [@it] |
| Expense claims and card spend | Example: Search for AI tool subscriptions | YYYY-MM-DD | [@finance] |
| Network and identity logs | Example: Sign-ins to known AI services | YYYY-MM-DD | [@security] |
| Product and engineering roadmaps | Example: Model registry and repository scan | YYYY-MM-DD | [@engineering] |
| Staff survey | Example: Short questionnaire per department | YYYY-MM-DD | [@ai-governance-lead] |
3. Register
| ID | System | Business purpose | Source | Our role (EU AI Act) | Legal classification (screening) | Internal risk tier | Business owner | Technical owner | Status | Last review | Next review |
|---|
| AIS-001 (Example) | Motor claims fraud flag | Flag claims for investigator review | In-house model | Provider and deployer | Minimal risk (motor insurance is outside Annex III point 5(c)) | High | [@claims-director] | [@ml-lead] | Live | YYYY-MM-DD | YYYY-MM-DD |
| AIS-002 (Example) | Life cover pricing model | Risk assessment and pricing for life insurance | In-house model | Provider and deployer | High-risk, Annex III point 5(c) | Critical | [@underwriting-director] | [@ml-lead] | In development | YYYY-MM-DD | YYYY-MM-DD |
| AIS-003 (Example) | Customer service chatbot | Answer policy questions | Third-party, built on a general-purpose AI model | Deployer | Article 50(1) disclosure, provider duty; deployer checks it is in place | Medium | [@service-director] | [@digital-lead] | Live | YYYY-MM-DD | YYYY-MM-DD |
| AIS-004 (Example) | CRV Shortlist 2.4 (Tarnwick Analytics) | Rank job applications | SaaS | Deployer | High-risk, Annex III point 4(a) | High | [@hr-director] | [@hr-systems] | Live | YYYY-MM-DD | YYYY-MM-DD |
| AIS-005 (Example) | Coding assistant | Code suggestions for developers | SaaS | Deployer | Minimal risk | Medium | [@cto] | [@platform-lead] | Live | YYYY-MM-DD | YYYY-MM-DD |
| AIS-006 (Example) | Email summary feature in office suite | Summarise long email threads | Embedded feature | Deployer | Minimal risk | Low | [@coo] | [@it] | Live | YYYY-MM-DD | YYYY-MM-DD |
| AIS-007 (Example) | Legacy document OCR | Read scanned claim forms | In-house | Provider and deployer | Minimal risk | Low | [@claims-director] | [@it] | Retired | YYYY-MM-DD | n/a |
Additional fields per system (keep in a linked record or extra columns):
| Field | Why it matters |
|---|
| Model or provider and version | Change tracking and incident response |
| Personal data processed (Y/N) and ROPA reference | Links to compliance/gdpr-ropa and compliance/dpia |
| Affected people | Customers, applicants, staff, public |
| Decision role | Informs, recommends, decides automatically |
| Human oversight arrangement | Links to ai-data/human-oversight-design |
| Jurisdictions of use | Drives the applicability register |
| Key documents | Intake, risk assessment, FRIA, model card, vendor questionnaire, instructions for use |
| Incident plan and contact | Links to ai-data/ai-incident-report |
| Contract and exit terms | Needed for decommissioning (section 7) |
4. Classification rules
4.1 Role
| Role | Rule of thumb | Follow-up |
|---|
| Provider | We develop the system or have it developed and place it on the market or put it into service under our name, including for internal use | ai-data/ai-act-technical-documentation, ai-data/ai-act-instructions-for-use where high-risk |
| Deployer | We use a system under our authority in a professional activity | ai-data/ai-act-deployer-obligations where high-risk |
| Both | In-house system used internally | Both sets of duties |
4.2 Legal classification (EU AI Act, as a dated example)
| Category | Screening question | Next step |
|---|
| Prohibited | Does it match an Article 5 practice? | Stop and escalate to legal |
| High-risk | Is it an Annex I safety component or an Annex III use case? | ai-data/ai-use-case-intake; if a provider believes Article 6(3) applies, ai-data/ai-act-high-risk-exemption-assessment |
| Transparency | Does Article 50 apply (chatbots, synthetic content, emotion recognition, deep fakes)? | ai-data/ai-content-transparency-plan |
| General-purpose AI model | Are we a provider of a general-purpose AI model? | ai-data/gpai-model-documentation, ai-data/gpai-copyright-policy |
| Minimal risk | None of the above | Usage policy and literacy only |
| Unclear | Not enough information | Owner to complete intake within 30 days |
Other jurisdictions record their own categories in the applicability
register. As dated examples: US federal agencies must inventory their
AI use cases at least yearly and publish a public version under OMB
Memorandum M-25-21 (3 April 2025); UK central government departments
publish records under the Algorithmic Transparency Recording Standard
(mandatory for departments and certain arm's-length bodies under the
December 2024 scope policy).
4.3 Internal risk tier
| Tier | Criteria (Example) | Review frequency |
|---|
| Critical | High-risk legal category, or automated decisions with legal effects on individuals | Every 6 months |
| High | Significant effect on customers, applicants or staff, or sensitive data | Every 12 months |
| Medium | External-facing or business-critical, limited effect on individuals | Every 12 months |
| Low | Internal productivity, no personal data beyond business contact data | Every 24 months |
5. Summary view
| Measure (Example) | Count |
|---|
| Systems registered | 7 |
| Live or in development | 6 |
| Retired | 1 |
| High-risk (EU screening) | 2 (AIS-002, AIS-004) |
| Article 50 transparency relevant | 1 (AIS-003) |
| Minimal risk | 4 (AIS-001, AIS-005, AIS-006, AIS-007) |
| Critical or High internal tier | 3 (AIS-001, AIS-002, AIS-004) |
| Reviews overdue | [Number] |
6. Maintenance and review
| Trigger | Action | Owner |
|---|
| New purchase or contract renewal with AI features | Add or update entry before signature | [@procurement] |
| New model, new release or change of intended use | Update entry and reclassify | [@technical-owner] |
| Vendor adds an AI feature to existing software | Add entry within 30 days | [@it] |
| Incident or complaint | Update risk tier if needed | [@business-owner] |
| Law or guidance change | Re-screen affected entries | [@legal] |
| Scheduled review date reached | Confirm all fields | [@business-owner] |
ISO/IEC 42001:2023 alignment (clause numbers only, original wording):
| Clause or control area | How this register supports it |
|---|
| 4.3 Scope of the AI management system | Lists the AI systems inside the scope |
| 6.1 Actions to address risks and opportunities (AI risk assessment and AI system impact assessment) | Gives the list of systems to assess |
| 7.5 Documented information | Register is controlled documented information |
| Annex A, A.4 Resources for AI systems | Records systems, tools and owners |
| Annex A, A.10 Third-party and customer relationships | Records supplier-provided systems |
7. Decommissioning
| Step | Detail (Example for AIS-007) |
|---|
| Decision and date | Retired YYYY-MM-DD, replaced by the claims platform OCR |
| Users informed | Claims team, YYYY-MM-DD |
| Data and logs | Logs kept for the retention period in the records schedule, then deleted |
| Contracts | Not applicable, in-house |
| Residual obligations | Keep technical documentation for the period required by law where applicable |
| Register | Status set to Retired; row kept for audit trail |
8. Glossary and references
| Term | Meaning |
|---|
| AI system | Machine-based system that infers from input how to generate outputs, as defined in the scope section |
| Provider | Develops a system and places it on the market or puts it into service under its own name |
| Deployer | Uses a system under its authority in a professional activity |
| Shadow AI | AI used for work without organisational approval |
| Internal risk tier | The organisation's own rating, separate from legal categories |
- NIST AI 100-1, Artificial Intelligence Risk Management Framework (AI RMF 1.0), January 2023, GOVERN 1.6 (inventory) and GOVERN 1.7 (decommissioning): https://www.nist.gov/itl/ai-risk-management-framework ; Playbook: https://airc.nist.gov/airmf-resources/playbook/govern/
- ISO/IEC 42001:2023, Information technology, Artificial intelligence, Management system. Official text: ISO, purchase or licensed access: https://www.iso.org/standard/81230.html
- Regulation (EU) 2024/1689 (AI Act), Articles 3, 5, 6, 25, 50, Annex III: https://eur-lex.europa.eu/eli/reg/2024/1689/oj ; amended by Regulation (EU) 2026/1744: http://data.europa.eu/eli/reg/2026/1744/oj
- OMB Memorandum M-25-21, 3 April 2025: https://www.whitehouse.gov/wp-content/uploads/2025/02/M-25-21-Accelerating-Federal-Use-of-AI-through-Innovation-Governance-and-Public-Trust.pdf
- UK Algorithmic Transparency Recording Standard hub: https://www.gov.uk/government/collections/algorithmic-transparency-recording-standard-hub
- Related templates: compliance/ai-regulatory-applicability-register, ai-data/ai-use-case-intake, ai-data/ai-usage-policy, ai-data/ai-literacy-plan, compliance/ai-vendor-risk-questionnaire, compliance/iso-42001-ai-policy, ai-data/ai-act-deployer-obligations
Standard and licence note
Structure aligned with NIST AI RMF 1.0 (a US Government publication,
GOVERN 1.6 and 1.7) and ISO/IEC 42001:2023 (a copyrighted standard
cited by clause and control area numbers only; no text reproduced), and
based on Regulation (EU) 2024/1689 as amended by Regulation (EU)
2026/1744 for the role and category columns. All wording is original.
Check the current official versions of each source before relying on
the classification rules. All organisations and systems in examples are
fictional. This is a documentation format, not legal advice. Template
licence: MIT.