Objectives
Purpose
This document establishes the measurable objectives for Rygen Technologies’ AI Management System (AIMS) in accordance with ISO/IEC 42001:2023. These objectives support the AI Policy and demonstrate our commitment to responsible, effective, and trustworthy AI deployment.
Scope
These objectives apply to all AI systems within the AIMS scope, including the X1 Integration Platform, Corsair TMS, and internal AI tools.
AIMS Objectives
Trustworthy Delivery
Objective: AI systems shall be evaluated for trustworthiness requirements prior to deployment, with explainability and human oversight controls implemented based on risk and impact assessment outcomes for mission-critical systems or decision-support applications.
Key Performance Indicators:
- Percentage of AI systems with completed trustworthiness evaluations before deployment
- Compliance with assessment-driven requirements in design reviews
Implementation:
- Conduct risk and impact assessments for all AI systems to determine trustworthiness requirements
- Include assessment-based explainability and oversight requirements in AI System Design templates
- Implement explainability and human oversight controls only where justified by risk assessment
- Enforce compliance checks during design reviews based on system-specific requirements
- Document rationale for systems not requiring explainability or human oversight controls
Responsibility: Principal AI Engineer
Review Frequency: Every four months
Performance Excellence
Objective: Achieve and maintain at least 90% accuracy/performance targets for all production AI systems as defined in their specifications.
Key Performance Indicators:
- Percentage of AI systems meeting defined performance targets
- Number of performance-related incidents
Implementation:
- Define specific performance targets in each AI System Design Document
- Measure performance from each production system’s evaluation results (AI-017)
- Conduct four-monthly performance reviews
- Address performance below target through the applicable AI-013 change or AI-023 incident process
Responsibility: Principal AI Engineer
Review Frequency: Every four months
Responsible Governance
Objective: Complete 100% of required AI risk assessments and impact assessments prior to production deployment.
Key Performance Indicators:
- Percentage of AI systems with completed risk assessments before deployment
- Percentage of AI systems with completed impact assessments before deployment
Implementation:
- Check at each management review that each system deployed in the period has a risk and impact assessment dated before deployment (AI-017)
- Maintain assessment templates in the aims-governance repository
- Train the teams that build AI systems on assessment requirements
Responsibility: Principal AI Engineer
Review Frequency: Every four months
Client Value Through Innovation
Objective: Deploy at least two new AI systems per year.
Key Performance Indicators:
- Number of new AI systems deployed annually
Implementation:
- Use client needs to prioritize AI features
- Request client feedback annually (AI-007)
Responsibility: Principal AI Engineer
Review Frequency: Every four months
Operational Resilience
Objective: Maintain 99% availability for AI services with documented fallback mechanisms for all critical AI features.
X1 AI Review is exempt from the 99% availability target.
Key Performance Indicators:
- AI service availability percentage
- Time to recovery for AI service incidents
Implementation:
- Monitor availability through GCP monitoring
- Document the manual process or other fallback for each critical feature
- Maintain incident response runbooks
- Conduct post-incident reviews to improve resilience
Responsibility: Principal AI Engineer
Review Frequency: Every four months
Compliance
Objective: Achieve 100% compliance with all relevant regulations and standards.
Key Performance Indicators:
- Number of compliance violations
- Percentage of compliance requirements met
- Audit findings and corrective actions completed
Implementation:
- Implement all controls per Statement of Applicability
- Address audit findings within defined timelines
- Maintain ISO 42001 certification through the DEKRA surveillance audits, the first due March 2027
Responsibility: Principal AI Engineer
Review Frequency: Every four months
Monitoring and Measurement
Each objective shall be:
- Measured using the defined KPIs
- Reviewed at the specified frequency
- Reported to top management at each management review
- Reviewed at governance committee meetings as needed
- Updated annually or when strategic changes occur
Integration with AIMS
These objectives:
- Support the AI Policy commitments
- Align with Rygen’s strategic direction
- Provide framework for risk assessment priorities
- Guide resource allocation decisions, which the CTO makes (AI-005)
- Inform training and competence requirements
Communication
These objectives shall be:
- Communicated to all relevant personnel
- Published in the AIMS Handbook on GitLab Pages (AI-006)
- Included in AIMS training materials
- Communicated during onboarding
- Discussed at AI Governance Committee meetings, which are held as needed
Revision History
| Version | Date | Author | Summary of Change |
|---|---|---|---|
| 1.0 | 2025-06-05 | Field Bradley | Initial draft. |
| 1.1 | 2025-09-02 | Field Bradley | Migrated to markdown and gitlab |
| 1.2 | 2025-10-08 | Field Bradley | Updated operational resiliency target from 99.9% to 99% |
| 1.3 | 2026-07-14 | Field Bradley | Objective review frequency and management-review reporting aligned to the every-four-months review cadence (AI-1759) |
| 1.4 | 2026-11-04 | Field Bradley | AI Governance Committee meetings are held as needed, not on a schedule (AI-2173). Removed the four-monthly fallback test and its KPI; the fallback for each critical feature is the manual process it automates, which needs no separate test (AI-2131) |
| 1.5 | 2026-11-09 | Field Bradley | Objectives are measured at each management review by the Principal AI Engineer as AI-017 sets, unmeasured KPIs are removed, X1 AI Review is exempt from the availability target, and Client Value counts at least two new AI systems a year (AI-2173, AI-2131; Field, 2026-11-07). The AI Policy link resolves on the published site. Removes the explainability and oversight percentage KPIs, and assessment training covers the teams that build AI systems. Performance response, prioritization, fallback and time-to-recovery wording aligned to existing processes. Sprinto compliance tracking removed. |