Roles and Responsibilities

Purpose

This document defines and communicates the roles, responsibilities and authorities for implementing, maintaining and continually improving the AI Management System (AIMS) under ISO/IEC 42001:2023.

It sets out who decides what, and how engineering and product teams share the implementation and maintenance of AI capabilities.

Scope

This document applies to every department that develops, deploys, governs or operates AI systems within the AIMS scope defined in AI-003: Scope.

Top-Level Responsibility and Oversight

Principal AI Engineer – Owner of the AIMS

The Principal AI Engineer owns the AIMS and is accountable for its establishment, implementation, maintenance and continual improvement, and for the AI systems within its scope. The role ensures that Rygen’s AI systems are developed and deployed in line with the ISO 42001 principles of ethical, transparent and trustworthy AI. Its responsibilities and authorities are:

  • Define and manage the AI governance framework.
  • Conduct risk and impact assessments for AI features, or oversee them when a Lead Product Manager leads them.
  • Oversee bias mitigation, explainability and data accountability practices.
  • Ensure ISO 42001 documentation and records are maintained.
  • Lead periodic reviews of the AIMS for improvement opportunities.
  • Work with product and infrastructure teams to build compliance into development workflows.
  • Give technical approval, before deployment, for new AI systems and for material changes affecting AI behavior, capabilities or risk.
  • Resolve whether a capability falls within the AI Engineering Team’s area of responsibility, applying the scope in AI-003.
  • Work with the CTO on high-risk decisions, exceptions, and matters that need broader engineering alignment.
  • Approve the decisions AI-006, AI-008, AI-009, AI-011, AI-013, AI-014, AI-015 and AI-023 assign to the Principal AI Engineer.

Technical approval complements the risk-based approvals required by AI-013: AI System Development Lifecycle.

Chief Technology Officer (CTO) – Executive Sponsor of the AIMS

The CTO is the executive sponsor and internal champion of the AIMS. All AIMS performance and progress are reported to the CTO. The CTO’s responsibilities are:

  • Provide strategic oversight and support for implementing and enforcing the AIMS.
  • Allocate resources for AI governance, infrastructure and compliance.
  • Act as the escalation point for AIMS concerns and incidents.
  • Approve new handbook documents, every revision of an AIMS policy and of AI-002, AI-003, AI-004, AI-005 and AI-019, and major revisions of other handbook documents (AI-006), and the decisions AI-008, AI-013 and AI-015 assign to the CTO.
  • Represent AI governance to the Board and external stakeholders when needed.

Chief Executive Officer (CEO)

The CEO:

  • Approves, together with the CTO, the acceptance of Critical risks and the deployment of Critical risk systems; this authority is not delegated (AI-008, AI-013).
  • Attends each management review and may call an ad hoc review (AI-012).

Cross-Functional Roles and Responsibilities

Team Leads

All technical team leads, including development, DevOps and data teams, ensure their teams operate in compliance with the AIMS. They:

  • Ensure their teams understand and apply AIMS policies and procedures.
  • Contribute to internal audits.
  • Report AI system performance, risk indicators and incidents to the Principal AI Engineer.
  • Support model documentation, risk assessments and feedback loops during development and operations.
  • Take part in AIMS reviews and continual improvement.

Information Security Officer

The Information Security Officer:

  • Owns AI-016 and reviews AI-016, AI-020 and their related procedures.
  • Evaluates and approves new AI tools, and maintains the Approved AI Tools List (AI-020.1) with each tool’s safe usage constraints (AI-020).
  • Assesses the security, privacy and compliance impact of AI incidents, decides whether regulatory notification is needed, and makes sure security controls are updated (AI-023).
  • Maintains security monitoring of AI infrastructure (AI-010).

AI Governance Committee

The Principal AI Engineer chairs the AI Governance Committee, and the CTO advises it. Its members and their roles are listed in AI-019.

The committee:

  • Reviews AI governance policies and frameworks before they are approved.
  • Reviews major AI systems for alignment with ethical and regulatory expectations.
  • Is informed of incident reports and audit findings in #ai-governance.
  • Advises on long-term AI governance strategy and the compliance roadmap.

Technical Responsibilities and Authorities

The AI Engineering Team is responsible for the technical direction, design, development and lifecycle oversight of AI systems.

AI Engineering Team

The AI Engineering Team is responsible and authorized to:

  • Own the AI-specific lifecycle: feasibility, impact and risk assessment, architecture, design, development, evaluation, deployment, monitoring, maintenance and improvement.
  • Take part early in planning and designing features and platforms that integrate with AI systems or expose capabilities to AI agents.
  • Set technical standards and requirements for AI systems in line with the AIMS.
  • Review new AI systems, and material changes to existing systems, before deployment.
  • Define and update model development, evaluation and validation protocols.
  • Propose updates to AIMS documentation and policies based on emerging practices or findings.
  • Recommend changes to AI features or workflows based on bias monitoring, risk assessments or performance anomalies.
  • Lead experimentation with new AI techniques and confirm they fit Rygen’s AI governance principles.

Product Teams

Product teams are responsible for:

  • Owning product requirements, business priorities and desired outcomes.
  • Working with the AI Engineering Team and other engineering teams to define scope and evaluate feasibility.
  • Defining requirements for human-in-the-loop override and safe fallback in AI-powered features.
  • Documenting intended use cases, assumptions and expected outcomes for AI or ML-enhanced features.
  • Working with the AI Engineering Team so that AI decisions are explainable and auditable from the product interface.

Lead Product Managers on the AI Governance Committee (AI-019) may lead a system’s problem definition (AI-024), impact assessment (AI-009) and risk assessment (AI-008). They do this with the AI Engineering Team, which takes part in each assessment and reviews it before approval. Approval follows AI-008 and AI-013.

Product teams do not decide AI architecture, technical implementation or governance requirements on their own.

Other Engineering Teams

Engineering teams that own platforms, infrastructure, APIs and integrations used by AI systems remain responsible for implementing, operating and maintaining them. They are responsible for:

  • Involving the AI Engineering Team before making architectural or implementation decisions that materially affect AI systems or agent-facing capabilities.
  • Taking part in planning, designing and reviewing AI systems that use their platforms.
  • Reviewing and approving, in the relevant GitLab merge request, AI system changes that materially affect their platforms, as required by AI-013: AI System Development Lifecycle.

The team-specific responsibilities below also apply.

Software Engineering Teams

Software engineering teams are responsible for:

  • Building AIMS-aligned practices into the codebase, including required logging, explainability hooks and feedback mechanisms.
  • Ensuring software modules that interface with AI components comply with AI governance controls.
  • Supporting the integration of AI-driven features under the validation and risk mitigation protocols set by the AI Engineering Team.
  • Taking part in peer reviews and technical audits of AI-related feature development and infrastructure changes.

Operational Systems Team

The Operational Systems Team is responsible for:

  • Implementing and maintaining the infrastructure for monitoring, logging and audit trails of AI model behavior and decisions.
  • Ensuring the deployment stack supports traceability, version control and rollback of AI models and their pipelines.
  • Managing data pipeline integrity and model-serving environments in a secure, reproducible and governed way.
  • Working with the AI Engineering Team to automate error handling and resource optimization for AI workloads.

Delegation and Shared Responsibility

The AI Engineering Team may delegate implementation and maintenance work to other engineering teams based on their expertise and platform ownership.

  • The teams involved agree on scope, responsibilities and technical requirements before work begins.
  • Work affecting AI behavior, capabilities or risk stays subject to the AI system lifecycle and its review and approval requirements.
  • Each engineering team stays responsible for the components it builds and operates.
  • The AI Engineering Team keeps accountability for AI-specific technical decisions and AIMS compliance, under the Principal AI Engineer’s overall accountability.

Delegation does not remove the need for early involvement, collaboration or technical review.

Working Principle

The AI Engineering Team has technical authority over Rygen’s AI systems. Other engineering teams keep ownership of their platforms and may build AI-related capabilities through collaboration or agreed delegation.

Teams must involve the AI Engineering Team before designing or deploying AI systems or making material changes to agent-facing capabilities.

Design is everything that follows ideation. An idea needs no set level of detail, and each team may record its ideas in its own way. Involving the AI Engineering Team can be as simple as notifying the team, which starts the AI system development lifecycle defined in AI-013: AI System Development Lifecycle.

Reporting Structure

At each management review (AI-012), the Principal AI Engineer reports to top management in the management review report on:

  • progress on AIMS milestones
  • nonconformities, incidents and stakeholder concerns
  • audit preparation and outcomes
  • training and awareness progress
  • recommendations for resource or process changes

Between management reviews, the Principal AI Engineer raises urgent matters with the CTO in Slack or 1:1 meetings. AIMS work is tracked in Jira.

The AI Governance Committee holds no regular meetings. It works in the #ai-governance Slack channel and meets as needed (AI-019). The committee:

  • confirms continued ISO 42001 alignment
  • reviews updates to policies and risk treatment; the authority named in AI-006, AI-008 or AI-013 approves the GitLab merge request

Communication and Awareness

All personnel involved with AI systems, including developers, analysts, QA and support, are told their responsibilities through:

  • AIMS onboarding documentation and checklists
  • internal training on responsible AI and ISO 42001 principles
  • AIMS steps in the development lifecycle, such as design reviews

Continuous Improvement

All teams contribute to the continual improvement of the AIMS by:

  • reporting issues, risks and improvement opportunities
  • taking part in retrospectives and governance debriefs
  • working across teams to refine compliance workflows and tooling

Revision History

VersionDateAuthorSummary of Change
1.02025-06-05Field BradleyInitial draft.
1.12025-09-02Field BradleyMigrated to markdown and gitlab
1.22026-07-14Field BradleyGovernance review cadence changed from quarterly to every four months (AI-1759)
2.02026-11-04Field BradleyLinked scope definitions to AI-003; clarified AI Engineering Team authority, platform-owner approval under AI-013, product responsibilities, and mandatory collaboration and delegation requirements; reset the next review date.
2.12026-11-09Field BradleyCut repeated text, records the approvals each procedure assigns to the CTO, CEO and Principal AI Engineer, defines the Information Security Officer, and removes drift detection (AI-2173, AI-2131). The committee is informed of incidents and audit findings, and team leads contribute to internal audits.
2.22026-11-11Field BradleyLead Product Managers on the AI Governance Committee may lead problem definitions, impact assessments and risk assessments with the AI Engineering Team; approval authority is unchanged (AI-2205)
2.32026-11-11Field BradleyClarified that design follows ideation and that notifying the AI Engineering Team starts the AI-013 lifecycle (AI-2180)