Building an AI System Inventory 2026: 8-Step Guide

Practitioner note: This article is practice-oriented compliance documentation, not legal advice. We are a compliance specialist, not a law firm. For legally binding information, consult a licensed attorney.

TL;DR

  • An AI inventory is a prerequisite for Art. 4 (AI Literacy) + Art. 26 (deployer obligations)
  • 8-step build: Discovery → classification → GPAI tagging → risk → FRIA trigger → provider/deployer → training → update
  • Minimum columns: 12 (tool name, provider, use case, GPAI yes/no, data class, Annex III, risk, FRIA, AI literacy, approval date, owner, update)
  • Update cycle: on change + annually
  • Fine exposure: up to EUR 15 million / 3 % (Art. 99(4) — high-risk/operator obligations); up to EUR 7.5 million / 1 % (Art. 99(5) — false information to authorities)

1. Why an AI inventory is mandatory

Even though the EU AI Act does not contain an explicit "inventory" article, an inventory is the operational minimum standard for several obligations:

2. 8-step guide

StepActivityOutput
1. DiscoveryEmployee survey, IT systems scan, licence reviewRaw list >=20 tools
2. AI system yes/noArt. 3 test per toolCleaned list
3. GPAI taggingModel classification (ChatGPT, Claude, Gemini = GPAI)GPAI column
4. Annex III check8 areas, review filter mechanismHigh-risk tag
5. FRIA triggerFor high-risk + certain deployers (Art. 27)FRIA to-do list
6. Provider/deployer roleDocument per toolRole column
7. AI literacy training per toolIndividual training obligationsTraining plan
8. Update cycleQuarterly review dateOwners, deadlines

3. Excel template columns (minimum standard)

  1. Tool name + version
  2. Provider (manufacturer)
  3. Use case (one short sentence)
  4. GPAI yes/no (model ID if known)
  5. Data categories processed
  6. Annex III area (1-8 or "none")
  7. Risk class (prohibited / high / limited / minimal)
  8. FRIA status (not required / open / completed)
  9. AI literacy training (date)
  10. Approved by / date
  11. Responsible business unit
  12. Last update + next review
Template as Excel + 12 pre-filled SME examples in the EU AI Act Kit.

4. 12 SME examples of typical AI tools (as of 04/2026)

ToolUse caseGPAIRiskObligation
ChatGPT EnterpriseGeneral text generationyeslimited (Art. 50 transparency)AUP, AI literacy
Microsoft 365 CopilotOffice assistanceyes (GPT-4)limitedAUP, data classification
Claude (Anthropic)Research, code supportyeslimitedAUP, AI literacy
HR recruiting software (with ML score)Candidate scoringnoHIGH (Annex III, 4)FRIA, bias test
Payroll AI (anomaly detection)Error detectionnominimaldocumentation only
DeepL ProTranslationno (NMT)minimaldocumentation only
GrammarlyLanguage optimisationnominimaldocumentation only
Zendesk AI (chatbot)Customer supportyes (GPAI-based)limited (Art. 50)AUP, bot disclosure
Otter.ai (transcription)Meeting transcriptionyes (Whisper-based)minimaldocumentation only
Sales forecast toolRevenue forecastingnominimaldocumentation only
Credit scoring softwareCredit decision recommendationvariableHIGH (Annex III, 5)FRIA, bias test
Spam filter (classic)Email filterno (Bayes)minimaldocumentation only

5. Update cycle + triggers

6. Audit preparation

  1. Inventory <6 months old
  2. Per tool: approval workflow traceable
  3. FRIAs in place for high-risk use cases from 02 August 2026 (Digital Omnibus proposal of 19 November 2025: postponement to 02 December 2027 — not yet adopted)
  4. AI literacy training logbook with knowledge quiz
  5. Acceptable Use Policy <12 months old
  6. Data breach procedure also documented for AI tools

Frequently asked questions

Is an AI inventory mandatory?
Indirectly yes: Art. 26 EU AI Act requires deployer documentation for high-risk AI, and Art. 4 AI Literacy requires knowledge of the AI systems used. Neither is feasible without an inventory.
Which tools count as an AI system under Art. 3?
Machine-learning models, logic/knowledge-based systems with inference function, statistical methods with optimisation. ChatGPT, Copilot, Claude, Gemini, Mistral and all GPAI-based tools are AI systems.
How deep should the inventory go?
Per AI system: vendor, use case, data categories, GPAI yes/no, risk class, FRIA status, owner and approval date. At least 8 columns.
Who maintains the inventory?
The controller (delegated by management) or an appointed AI officer. Recommendation: central function, decentralised inputs from departments.
How often to update?
On every change of AI usage plus at least annually. Triggers: new tools, contract renewals, Annex III changes, trilogue adoption of the Digital Omnibus proposal (19 Nov 2025; not yet adopted).
Cost-effective starting point?
Excel plus the Compliance-Kit AI Inventory template (12 pre-filled SME examples). Total roughly 2-4 person-days for an SME of 50-100 employees.

Sources

As of: 02 May 2026

Tools & self-tests

EU AI Act Quick Test Classifies your AI system by risk class (Art. 6, Annex III). Fining Calculator Estimate the potential fine exposure for your organisation. EU AI Act Self-Assessment Classification and obligation mapping for every AI system in your organisation. AI Inventory Quick Check Systematic capture of your AI applications in 8 steps.