AI Assurance Institute Logo AI Assurance Institute

Our AI QMS Design

A System to manage Regulatory Compliance with the EU AI Act

The AI Quality Management System (AI QMS) has been purposefully designed and engineered from the ground up to ensure that high-risk AI systems are developed, deployed, and monitored in a way that actively safeguards the health, safety, and fundamental rights of all affected persons.

This is not an add-on or afterthought. Protection is embedded into the core architecture, workflows, and controls of the platform. Every component, process, and record is aligned with the requirements of the EU AI Act-particularly the obligations for risk management, post-market monitoring (Article 72), serious incident reporting (Article 73), and fundamental rights impact assessments-and the quality management system principles outlined in prEN 18286.

Five Pervasive QMS Processes: The Foundation of Protection

At the heart of the system are five interconnected processes that apply universally across all AI system activities. These processes ensure consistent, auditable protection:

  1. Traceability
    Every action, decision, data flow, and outcome is fully traceable back to the specific AI system, its context, the responsible parties, and the governing controls in place at the time. Rich contextual selectors capture AI System ID, reference, and purpose at the moment of interaction, creating an unbreakable chain of accountability.
  2. Review and Approval / Sign-off
    Structured workflows require appropriate levels of review and formal sign-off before critical activities proceed. Digital signing mechanisms (with timestamps and cryptographic hashes) provide verifiable evidence of approval, locking records against unauthorized changes.
  3. Evidence Generation and Record Keeping
    High-quality, immutable evidence is automatically generated and preserved as part of normal operations. Snapshots of entire governance structures, audit programmes, risk assessments, and control configurations are taken at key moments, ensuring that historical evidence remains intact even if master templates evolve.
  4. Nonconformity and Corrective Action
    The system proactively identifies deviations, risks, or incidents that could impact health, safety, or rights. Integrated mechanisms for logging nonconformities, initiating corrective and preventive actions (CAPA), and tracking them to closure ensure timely resolution and continuous improvement.
  5. Change Management
    All modifications-whether to AI systems, their configurations, governance frameworks, or operational parameters-are subject to rigorous, documented change control. Administrative and technical changes are themselves logged as controlled QMS events, creating a complete audit trail of how protection measures are maintained over time.

These five processes are reinforced by a Universal QMS Control Footer (or Control Bar) that appears on every controlled document, template, form, record, and dashboard. This standardized element ensures that users are always aware of traceability links, pending approvals, linked evidence, open nonconformities, and required change actions.

How the System is Designed to Meet Regulatory Requirements

EU AI Act Alignment (Focus on Protection of Health, Safety & Fundamental Rights)

EN 18286 Alignment (AI Quality Management Systems)

The platform implements a quality management system specifically tailored for AI, covering leadership, planning, support processes, operational controls, performance evaluation, and improvement.

Built-In Architectural Safeguards for Protection

Delivering Confidence and Compliance

By design, the AI Quality Management System provides organizations with demonstrable compliance with the EU AI Act, alignment with prEN 18286, audit-ready evidence, and operational tools that make protection practical and sustainable.

The result is an AI Quality Management System that does not merely document compliance but actively works to prevent harm and uphold the rights of affected persons-today and as AI technologies continue to advance.