Procedure
AI Assurance Case Procedure
This procedure defines how an assurance case for an AI system is established, argued, evidenced, reviewed, reported, maintained, and closed. ai/framework-ai-assurance-evidence.md owns the evidence model that the assurance case organizes: the claim taxonomy, the evidence classes, the attestation tiers, and the sufficiency criteria. That framework deliberately defines no assurance activity; this procedure supplies the activity. It adapts the assurance case structure of ISO/IEC/IEEE 15026-2:2022 (Clause 5) to the corpus's existing evidence model, by paraphrase and clause reference only: the 15026 series is a licensed proprietary reference and is not reproduced here.
- Type
- Procedure
- Version
- 1.0.6
- Date
- 2026-10-27
- Owner
- Chief Information Security Officer
- Approving authority
- Governance Library Maintainer
- Classification
- Public
- Confidentiality
- Public
- Category
- AI Governance
- Review frequency
- 6 to 12 months and upon material AI governance or regulatory change
- Licence
- CC BY-SA 4.0
Read the full text on GitHub (opens in a new tab)
Outline
- Purpose
- Scope
- Roles and responsibilities
- 1. Triggers
- 2. Establish the case
- 3. Develop the argument
- 4. Link evidence
- 5. Review
- 6. Report and gate submission
- 7. Maintenance and change
- 8. Retirement
- 9. Records
- Framework alignment
- Limitations
Frameworks it maps to
- ISO/IEC/IEEE 15026-1:2025
- ISO/IEC/IEEE 15026-2:2022
- ISO/IEC/IEEE 15026-4:2021
- NIST AI 100-1 (AI RMF 1.0)
- EU AI Act (Regulation (EU) 2024/1689 as amended by Regulation (EU) 2026/1744)
- International AI Safety Report 2026
Related documents
- AI Assurance Evidence and Attestation Frameworkai
- AI Assurance Case Templateai
- AI Governance and Risk Frameworkai
- AI System Audit and Certification Frameworkai
- AI Audit Procedureai
- AI Evaluation Procedureai
- AI Governance Council Charterai
- Audit Evidence Package Templatecompliance
Next
All Documents