Operational Quality Standard
Standard ID: GCAIS-STD-005Table of Contents
1. Scope
This standard establishes minimum requirements for the operational reliability, performance documentation, and incident management practices of organizations providing AI services in production. It applies to AI systems offered as services to external users or integrated into business-critical processes.
The requirements address the documented commitments an organization makes regarding service availability and performance, the processes in place to manage service degradation and failure, and the controls applied to system changes. Requirements apply to the organization responsible for operating the service, including where service components are provided by third parties.
Internal-use AI systems not offered as services and research or prototype deployments are excluded from the scope of this standard.
2. Definitions
- Service Level Agreement (SLA)
- A documented commitment specifying the expected availability, performance, and error rate thresholds for an AI service, and the remedies available to users when commitments are not met.
- Incident
- Any unplanned interruption to or degradation of an AI service, or any event that causes or risks causing a material deviation from committed service levels.
- Change Management
- The process by which modifications to AI systems or their supporting infrastructure are assessed, approved, tested, and deployed in a controlled manner.
3. Requirements
4. Assessment Criteria
Assessment includes review of published SLA documents, examination of monitoring and alerting infrastructure, review of incident response records, change management records for the preceding 12 months, and inspection of dependency documentation currency.
5. Compliance Indicators
| Requirement | Compliant Indicator | Non-Compliant Indicator |
|---|---|---|
| REQ-005-1 | SLA published, current, and accessible to users | No SLA, or SLA lacks availability targets or error rate commitments |
| REQ-005-2 | Error rate thresholds documented and monitoring in place | No thresholds, or no monitoring of error rates in production |
| REQ-005-3 | Incident response plan current; post-incident reviews documented | No plan, or incidents not reviewed within required timeframes |
| REQ-005-4 | Change records show pre-deployment testing for material changes | No change procedure, or material changes deployed without testing records |
| REQ-005-5 | Dependency documentation current and updated within required intervals | No documentation, or documentation not updated following known changes |
6. Version History
| Version | Date | Notes |
|---|---|---|
| v1.0 | 2026-03-01 | Initial publication. Adopted by Standards Board resolution 2026-02-26. |