The European Commission's October 19 announcement identifies EN 18286:2026 as the first harmonised standard supporting the AI Act. Its focus is the quality management system required by Article 17 for providers of high-risk AI systems.
The important limit
The Commission says implementing the standard gives a presumption of conformity with the Article 17 requirements covered by it. That is narrower than compliance with every obligation in the AI Act. Do not describe a quality-management standard as a complete legal clearance for an AI product.
Build on existing management systems
The announcement says EN 18286 is organized around the AI system lifecycle and can be layered on existing standards such as ISO/IEC 42001, ISO 9001, or sector-specific systems. For an established organization, the practical work may be mapping and extending current processes rather than creating another parallel bureaucracy.
Identify how your existing system handles requirements, development changes, testing, suppliers, incidents, and post-deployment monitoring. Then determine where AI-specific evidence or responsibilities are missing.
Before buying certification services
Confirm that your organization is a provider of a system in scope and determine the obligations and dates applicable to that system. Ask advisers which requirements their proposed work addresses and which remain outside it.
This is a meaningful implementation milestone, but not a reason to market a product as universally compliant. The useful deliverable is a documented, operating quality system supported by evidence, with an explicit map from each control to the obligation it is intended to satisfy.
Sources & further reading
Reviewed October 26, 2026. Product availability and regulatory guidance can change. Vendor performance claims are not independent Rovio AI benchmarks.
Found this useful? Share it.