Statement of Applicability

1. Purpose

This Statement of Applicability (SOA) defines which ISO 42001:2023 Annex A controls are applicable to Rygen’s AI Management System (AIMS). It provides:

  • A complete listing of all Annex A controls
  • The applicability determination for each control (Included/Excluded)
  • Implementation status for included controls
  • Justification for any excluded controls
  • Evidence references demonstrating control implementation

This document fulfils the requirements of ISO 42001:2023 clause 6.1.3f and serves as the formal controlled record for audit purposes.

2. Scope

This SOA covers the AIMS scope as defined in AI-003 (AIMS Scope), which includes:

  • All AI systems developed and operated by Rygen
  • AI systems provided to clients through Rygen’s products and services
  • Third-party AI models and services used inside Rygen’s AI systems, as components of those systems (AI-003, What Counts as an AI System)

3. SOA Methodology

Applicability decisions are made using a risk-based approach aligned with the AI-008 Risk Management Framework:

  1. Risk Assessment: Each control is evaluated against identified AI risks
  2. Business Context: Operational requirements and regulatory obligations are considered
  3. Proportionality: Control implementation is proportionate to the risk level
  4. Evidence Requirements: Each included control requires documented evidence of implementation

Controls may only be excluded where:

  • The control addresses risks not present in Rygen’s context
  • Alternative controls adequately address the underlying risk
  • Regulatory requirements do not mandate the control

All exclusion decisions are documented with explicit justification.

Note: Treatment measures that implement these controls are documented in the Risk Treatment Register (treatments.yaml) within the AIMS Governance repository and in individual risk reports. These are operational measures, not additional control objectives.

4. Annex A Controls

Control IDControl NameApplicableStatusJustification/Evidence
A.2.2AI policyYesActiveAI Policy document (AI-001), version-controlled in Gitlab.
A.2.3Alignment with other organizational policiesYesActiveAI Policy (AI-001), Policy Alignment section; merge request review of policy changes checks consistency with organizational policies.
A.2.4Review of the AI policyYesActiveAI Policy (AI-001), Review section: the policy is reviewed at each management review as a standing AI-012 input; policy revisions are approved in GitLab merge requests. The September 2026 management review report records, by a correction dated November 6, 2026 on the AIMS owner’s statement, that the policy was reviewed at the September 21, 2026 meeting and remains suitable, with no change required; the minutes hold no record of it. The November 2026 handbook periodic review (AI-1976) also reviewed the policy.

4.2 Internal Organization (A.3)

Control IDControl NameApplicableStatusJustification/Evidence
A.3.2AI roles and responsibilitiesYesActiveRoles, Responsibilities, and Authorities document (AI-005) in Gitlab, defining specific AIMS roles.
A.3.3Reporting of concernsYesActiveFeedback routes in AI-007, Feedback Channels section. Internal concerns are raised as the Jira ‘Concern’ issue type in the AI project (for example AI-1289). Client concerns are HubSpot Service tickets labeled under the Customer Success AI-Related Request SOP (Confluence) and escalated to the AIMS owner. Concerns raised during development are recorded in impact assessments (AI-009).

4.3 Resources for AI Systems (A.4)

Control IDControl NameApplicableStatusJustification/Evidence
A.4.2Resource documentationYesActiveSystem Design Specifications identify each system’s personnel, technical and infrastructure resources (AI-006 Resource Documentation), in a System Resources section or, for GOV-086 and GOV-093, across their architecture, provider and data sections, with personnel in GOV-086 Section 10.5 and in GOV-093’s access controls and support ownership (Sections 5.3 and 10.4); AI-005 defines AIMS-level roles and responsibilities.
A.4.3Data resourcesYesActiveAI Data Management Process (AI-011); System Design Specifications document data handling for API-based systems.
A.4.4Tooling resourcesYesActiveSystem Design Specifications identify AI and machine learning, infrastructure and development tooling per system, in a System Resources section or, for GOV-086 and GOV-093, across their architecture, provider and data sections.
A.4.5System and computing resourcesYesActiveSystem Design Specifications with infrastructure sections; GCP platform monitoring of the services; Shipment Tracking Copilot workflow job recovery and redundant user interface (AC.48).
A.4.6Human resourcesYesActiveSystem Design Specifications identify the roles each system needs; AI role job descriptions in content/competence/. Employee AIMS acknowledgment and training records (AI-016, AI-020, AI-021) maintained in Leapsome. Contractors are not in Leapsome and acknowledge AI-016 and AI-020 through a Microsoft 365 form, responses held in M365 and visible to team leads (CAR-010).

4.4 Assessing Impacts of AI Systems (A.5)

Control IDControl NameApplicableStatusJustification/Evidence
A.5.2AI system impact assessment processYesActiveAI System Impact Assessment Process document (AI-009), version-controlled in Gitlab.
A.5.3Documentation of AI system impact assessmentsYesActiveCompleted AI System Impact Assessment reports for each system, using the standard template.
A.5.4Assessing AI system impact on individuals or groups of individualsYesActiveAI System Impact Assessment Process, Individual Impacts subsection (under Impact Categories); completed assessments with analysis of individual impacts.
A.5.5Assessing societal impacts of AI systemsYesActiveCompleted AI System Impact Assessments, Section 5 (Societal Impacts) and Section 5.2 (Safety Assessment).

4.5 AI System Life Cycle - Management Guidance (A.6.1)

Control IDControl NameApplicableStatusJustification/Evidence
A.6.1.2Objectives for responsible development of AI systemYesActiveAIMS Objectives document (AI-002), defining six measurable objectives and their KPIs.
A.6.1.3Processes for responsible AI system design and developmentYesActiveAI System Development Lifecycle (AI-013), defining the seven-phase lifecycle.

4.6 AI System Life Cycle - Implementation (A.6.2)

Control IDControl NameApplicableStatusJustification/Evidence
A.6.2.2AI system requirements and specificationYesActiveAI System Design Specifications, a Phase 1 deliverable under AI-013.
A.6.2.3Documentation of AI system design and developmentYesActiveDeliverables required by AI-013, including design specifications, risk assessments, decision records and evaluation reports.
A.6.2.4AI system verification and validationYesActivePhase 4 evaluation results under AI-013, against each system’s acceptance criteria: evaluation reports (EVAL-) in the aims-governance repository, or results in the system repository or MLflow linked from the Jira ticket or merge request; user acceptance results where conducted.
A.6.2.5AI system deploymentYesActiveThe deployment plan recorded or referenced in each system design specification (AI-013 Phase 5); Corsair Suite Branching and Release procedure and Application Release Gateway (Confluence); GitLab continuous integration and deployment pipeline whose automated tests must pass before production; release tickets in the Jira DevOps project recording QA review and product manager sign-off. Confirmed at the February 2026 internal audit (AUD-2026-02).
A.6.2.6AI system operation and monitoringYesActiveEach AI system’s monitoring plan (MON-001, MON-002, MON-003, MON-004, MON-007, MON-008, MON-009, MON-010, MON-011) requires results at each management review (AI-012, AI-017). The September 2026 review records that no four-monthly monitoring report was held for any system (Section 2.3); the evidence held is that review’s AIMS objectives, GCP availability and AI Service Provider Performance results, the monthly model-deprecation review (AC.72) and incident record CAR-011.
A.6.2.7AI system technical documentationYesActiveFor Rygen’s engineering, support and audit, the System Design Specification for each AI system, version-controlled in GitLab as AI-013 requires; for users, the Corsair and X1 user documentation sites (A.8.2).
A.6.2.8AI system recording of event logsYesActiveSystem logs and DevOps runbooks defining logging requirements, both stored in Google Cloud (GCP); production logs are kept for 365 days in Rygen’s production project and 30 days in the client X1 projects (each system design specification).

4.7 Data for AI Systems (A.7)

Control IDControl NameApplicableStatusJustification/Evidence
A.7.2Data for development and enhancement of AI systemYesActiveAI Data Management Process (AI-011); System Design Specifications with Data Management sections for API-based systems.
A.7.3Acquisition of dataYesActiveSystem Design Specifications document data sources, acquisition methods, and integration points per AI-011.
A.7.4Quality of data for AI systemsYesActiveSystem Design Specifications include data quality assessments for API-based systems per AI-011.
A.7.5Data provenanceYesActiveSystem Design Specifications document data flows, transformation pipelines, and audit trails per AI-011; pipeline code in application source repositories.
A.7.6Data preparationYesActiveSystem Design Specifications document data transformation and preparation pipelines per AI-011; preparation code in application source repositories.

4.8 Information for Interested Parties (A.8)

Control IDControl NameApplicableStatusJustification/Evidence
A.8.2System documentation and information for usersYesActiveCorsair and X1 user documentation sites; live training and onboarding delivered by Customer Success during implementation; in-app notifications and disclaimers. Confirmed at the February 2026 internal audit (AUD-2026-02).
A.8.3External reportingYesActiveAI-007 Communication Policy, Feedback Channels section (External Channels: ai-admin@rygen.com and support tickets) and External Communications section (End Users: how to report issues, via in-app notifications and the support portal). Client reports are HubSpot Service tickets labeled under the Customer Success AI-Related Request SOP (Confluence) and escalated to the AIMS owner, as for A.3.3.
A.8.4Communication of incidentsYesActiveAI Incident Response Procedure (AI-023), Communication Requirements section: communication follows the Engineering Incident Response Procedure, with AIMS recipients and channels per AI-007 (Communication Triggers, AI Incidents); incident records CAR-001, CAR-002 and CAR-011. CAR-011 records that the affected clients were notified by their account manager; CAR-001 and CAR-002 show no notification within the earlier AI-007 timelines.
A.8.5Information for interested partiesYesActiveImpact assessment communication plans within individual system impact assessments (Section 8); AI-007 Communication Policy, Communication Framework section; records of communications with clients, auditors, and partners.

4.9 Use of AI Systems (A.9)

Control IDControl NameApplicableStatusJustification/Evidence
A.9.2Processes for responsible use of AI systemsYesActiveAI Policy (AI-001), Policy Statement section (Responsible AI and Trustworthy Delivery subsections); AI System Development Lifecycle (AI-013), which embeds responsible use controls.
A.9.3Objectives for responsible use of AI systemYesActiveAIMS Objectives document (AI-002), specifically objectives for Trustworthy Delivery (explainability, human oversight).
A.9.4Intended use of the AI systemYesActiveCompleted AI System Impact Assessments (Section 2.1 Intended Uses, Section 2.2 Foreseeable Misuses) and AI System Design Specifications.

4.10 Third-party and Customer Relationships (A.10)

Control IDControl NameApplicableStatusJustification/Evidence
A.10.2Allocating responsibilitiesYesActiveThird-Party Provider Dependencies section of each System Design Specification, recording each provider’s role and data-handling terms; AI providers are used under their standard API terms, with the paid tier for Gemini (R026); provider terms are kept in Sprinto vendor records (AI-006).
A.10.3SuppliersYesActiveSupplier due diligence records (SOC2 reports, privacy terms, service agreements) in Sprinto; provider performance summarized for each management review (AI-012 Preparation) and recorded in the review report’s AI Service Provider Performance appendix; monthly model-deprecation review (AC.72).
A.10.4CustomersYesActiveInterested Parties analysis (GOV-061), Clients section; client requirements documented in implementation project plans and impact assessments.

5. Exclusions Summary

No controls are currently excluded from the scope of Rygen’s AIMS.

All 38 ISO 42001:2023 Annex A controls have been determined to be applicable based on Rygen’s risk profile, operational context, and commitment to comprehensive AI governance.

Exclusion Review History

DateControlDecisionRationale
2026-01-13A.8.3 (External reporting)Changed to IncludedPreviously excluded due to no regulatory reporting obligations. Re-assessed following pre-registration audit; AI-007 Communication Policy provides adequate implementation evidence for external reporting to auditors, clients, and compliance stakeholders.

6. Summary Statistics

CategoryCount
Total Annex A Controls38
Included Controls38
Excluded Controls0
Active Controls38
Planned Controls0
  • AI-001: AI Policy
  • AI-002: AIMS Objectives
  • AI-003: AIMS Scope
  • AI-005: Roles, Responsibilities, and Authorities
  • AI-007: Communication Policy
  • AI-008: Risk Management Framework
  • AI-009: AI System Impact Assessment Process
  • AI-011: AI Data Management Procedure
  • AI-013: AI System Development Lifecycle

8. Revision History

VersionDateAuthorSummary of Change
1.02026-01-13Field BradleyInitial controlled document created following pre-registration audit findings. Control A.8.3 included with AI-007 as evidence.
1.12026-01-13Field BradleyUpdated A.5.5 evidence to reflect safety impact assessment per ISO 42001 clauses 6.1.4 and A.5.5.
1.22026-02-04Field BradleyAdded clarifying note in Section 3 about Risk Treatment Register relationship to SoA controls (AI-1216)
1.32026-02-17Field BradleyUpdated evidence statements for A.2.3, A.4.2–A.4.6, A.7.2–A.7.6, and A.10.2 to reflect actual implementation evidence (design spec sections, AI-011 v1.2 applicability by system type)
1.42026-07-14Field BradleyA.2.4 evidence updated: AI policy review schedule is now every four months (AI-1759)
1.52026-11-09Field BradleyControl evidence matches data/annex_a.yaml, with front matter, scope, provider terms, log retention, monitoring plans and lifecycle evidence matched to their sources (AI-2173, AI-2131). Exclusions are no longer reviewed at each management review. A.8.4 evidence aligned to the Engineering Incident Response Procedure route.