---
layout: raw-data.njk
---

# Stage 7: Structure Proposal Evaluation Against AI Guidance

## Date and Agent
- Date: 2025-11-19
- Agent: general-purpose (opus)
- Scope: Final validation of Stage 6 structure proposal

## Executive Summary

The Stage 6 structure proposal presents a **single consolidated document** organized around the **conformance workflow**, with 9 major sections that guide users from understanding conformance requirements through risk assessment, standard implementation, and conformance demonstration. This evaluation against government AI guidance principles finds the structure **strongly aligned** with transparency, accessibility, and user-centered design principles, though the AI guidance collection provides limited direct validation since it focuses on AI systems rather than general documentation standards.

The proposed structure demonstrates **exceptional alignment** with accessibility requirements, transparency principles, and accountability frameworks found in the AI guidance. The single document approach eliminates navigation barriers identified in current fragmented presentation, while the conformance-centered organization reflects the transparency and process documentation principles emphasized throughout AI governance guidance. The structure is **ready to proceed to Phase 2** with high confidence, requiring only minor adjustments to strengthen alignment with progressive disclosure principles and ensure optimal accessibility for all users.

Critical success factors for Phase 2 include maintaining clear visual distinction between normative requirements and guidance, implementing effective navigation aids that support both linear and reference usage patterns, and ensuring all content remains accessible and findable through multiple pathways.

## Structure Proposal Overview

The Stage 6 proposal transforms 30 fragmented documents into a **single consolidated identification conformance resource** with:

- **9 major sections** organized around the conformance workflow (assess → select → implement → demonstrate)
- **Integrated standards and guidance** presented together with visual distinction
- **Role-based entry points** for different user types (implementers, assessors, auditors)
- **No hidden content** - elimination of all detail expanders
- **Conformance as the central organizing principle** rather than conceptual topics
- **Active voice guidance** while preserving core standards text
- **Essential additions**: Biometric privacy requirements (2-3 pages) and NCSC cybersecurity cross-references (1 page)

The structure addresses Tom's primary concern that conformance is "the whole point" yet currently "tucked away," by making conformance the literal first section and organizing all content to support the conformance journey.

## AI Guidance Queries Performed

**Important Context**: The generative-ai-guidance-gcdo MCP server contains AI/GenAI-specific guidance for responsible AI use in government, NOT general content design guidance. Evaluation focuses on extracting relevant principles that can be analogously applied.

Semantic searches performed:
1. "organizing complex technical documentation" - 10 results focusing on implementation guides
2. "single document vs multiple documents user navigation" - 10 results on transparency principles
3. "user-centered content organization workflow-based information architecture" - 10 results on framework alignment
4. "making content accessible and findable navigation in technical documents" - 10 results emphasizing accessibility
5. "presenting compliance requirements technical standards usability" - 10 results on mandated standards
6. "integrating requirements and guidance progressive disclosure techniques" - 10 results on guidance alignment
7. "transparency about processes auditing assessment requirements" - 10 results on transparency/accountability

## Evaluation Against AI Guidance Principles

### Overall Approach Validation

**Single consolidated document**:
The AI guidance consistently emphasizes reducing barriers to access and understanding. While not explicitly addressing single vs. multiple documents, the accessibility guidance states that agencies must "Build GenAI to mandated accessibility standards and test it with disabled people" ([DocRef](https://docref.digital.govt.nz/nz/generative-ai-guidance-gcdo/accessibility-and-genai/2025/en/#part3-subpart2)). A single document eliminates navigation barriers between 30 separate resources, directly supporting accessibility mandates.

**Validation**: STRONGLY SUPPORTED - Consolidation removes access barriers and improves findability.

**Conformance-centered organization**:
AI guidance emphasizes transparency about "development and use of AI... what stage the project is at, the intent of use or the problem it's trying to solve" ([DocRef](https://docref.digital.govt.nz/nz/generative-ai-guidance-gcdo/governance-and-genai-in-the-public-service/2025/en/#part2-det5)). Similarly, organizing identification standards around conformance stages (assess, implement, document, get assessed) reflects this principle of process transparency.

**Validation**: SUPPORTED BY ANALOGY - Process-centered organization aligns with transparency principles.

**9-section structure**:
The AI Framework's vision emphasizes "efficient, responsive and high-performing Public Services" ([DocRef](https://docref.digital.govt.nz/nz/generative-ai-guidance-gcdo/public-service-ai-framework/2025/en/#part2-para1)). The 9-section structure provides comprehensive coverage while maintaining navigability, supporting efficient service delivery.

**Validation**: APPROPRIATE - Depth and breadth support comprehensive yet navigable resource.

### Key Structural Elements Validation

**Integrated standards-guidance**:
AI guidance models integrated presentation throughout - technical requirements are consistently presented alongside implementation guidance rather than separated. The accountability principle requires "application of relevant regulatory and governance frameworks, reporting, auditing and/or independent reviews" ([DocRef](https://docref.digital.govt.nz/nz/generative-ai-guidance-gcdo/public-service-ai-framework/2025/en/#part3-det5-para2)) as a unified approach.

**Validation**: STRONGLY SUPPORTED - Integration reflects best practice in AI guidance presentation.

**Workflow-based order**:
The AI guidance emphasizes building "accountability and responsibility" through clear processes and controls ([DocRef](https://docref.digital.govt.nz/nz/generative-ai-guidance-gcdo/accountability-responsibility-and-genai/2025/en/#part4-subpart1-para1)). Workflow-based ordering creates clear process accountability.

**Validation**: WELL ALIGNED - Sequential workflow supports process understanding and accountability.

**No hidden content**:
Multiple AI guidance references emphasize accessibility as fundamental, linking repeatedly to accessibility standards ([DocRef](https://docref.digital.govt.nz/nz/generative-ai-guidance-gcdo/transparency-and-genai/2025/en/#part4-para1-2)). Hidden content violates accessibility principles by creating barriers for screen readers and cognitive accessibility.

**Validation**: MANDATORY ALIGNMENT - Eliminating hidden content is essential for accessibility compliance.

**Role-based entry points**:
AI guidance recognizes diverse audiences, emphasizing skills building "to ensure they understand how to use AI in line with organisational and public service values" ([DocRef](https://docref.digital.govt.nz/nz/generative-ai-guidance-gcdo/public-service-ai-framework/2025/en/#part6-para3-3)). Multiple entry points serve different skill levels and roles.

**Validation**: SUPPORTED - Multi-audience approach reflects AI guidance's recognition of diverse users.

### Navigation Strategy Validation

**Proposed navigation aids**:
The emphasis on transparency and accessibility throughout AI guidance ([DocRef](https://docref.digital.govt.nz/nz/generative-ai-guidance-gcdo/genai-and-customer-experience-with-government/2025/en/#part1-title)) supports comprehensive navigation aids including TOC, breadcrumbs, and cross-references.

**Validation**: WELL ALIGNED - Multiple navigation pathways support accessibility and findability.

**Cross-linking approach**:
Transparency principles require clear connections between related requirements and guidance. The proposed extensive cross-linking supports this transparency mandate.

**Validation**: SUPPORTED - Cross-references enhance transparency and understanding.

**Progressive disclosure**:
While AI guidance doesn't explicitly address progressive disclosure, the structure's approach of revealing information through the conformance workflow (assess → select → implement → demonstrate) aligns with logical information revelation.

**Validation**: IMPLICITLY SUPPORTED - Workflow progression naturally implements progressive disclosure.

## Strengths Identified

1. **Exceptional Accessibility Alignment**: The elimination of hidden content and single document approach strongly aligns with accessibility mandates emphasized throughout AI guidance, particularly the requirement to "Build GenAI to mandated accessibility standards" ([DocRef](https://docref.digital.govt.nz/nz/generative-ai-guidance-gcdo/accessibility-and-genai/2025/en/#part3-subpart2)).

2. **Strong Process Transparency**: The conformance-centered organization exemplifies the transparency principles that pervade AI guidance, making the assessment and compliance process clear and navigable rather than obscured.

3. **Integrated Presentation Excellence**: The integration of standards with implementation guidance reflects the AI guidance's own approach of presenting requirements alongside practical implementation, avoiding the separation that creates navigation burden.

4. **Multi-Pathway Navigation**: The combination of role-based entry points, comprehensive TOC, cross-references, and workflow progression provides multiple ways to access content, supporting diverse user needs and accessibility requirements.

5. **Clear Accountability Framework**: The structure makes conformance requirements and processes explicit, supporting the accountability and auditing requirements emphasized in AI governance guidance.

## Risks and Concerns

1. **Visual Distinction Challenge**: While the structure proposes visual distinction between normative standards and guidance, Phase 2 must ensure this distinction is clear enough to prevent confusion while maintaining integration benefits. The AI guidance doesn't provide specific patterns for this distinction.

2. **Document Length Concern**: A single consolidated document may become lengthy. While this improves navigation compared to 30 documents, Phase 2 should consider pagination strategies or modular presentation to prevent overwhelming users.

3. **Progressive Disclosure Depth**: The structure could benefit from more explicit progressive disclosure within sections. While the workflow provides macro-level progression, micro-level information revelation within sections needs attention.

4. **Search Optimization**: With all content in one document, search functionality becomes critical. The structure should plan for robust search capabilities with filtering options (requirements vs. guidance vs. examples).

## Refinement Recommendations

1. **Strengthen Progressive Disclosure**: Within each major section, implement clear progression from overview → key concepts → detailed requirements → implementation guidance → examples. This micro-level progression would complement the macro-level workflow progression.

2. **Enhance Visual Hierarchy Plan**: Develop specific visual patterns for distinguishing:
   - Normative requirements (e.g., bordered boxes with icons)
   - Implementation guidance (e.g., different background color)
   - Examples and case studies (e.g., indented with markers)
   - Cross-references (e.g., consistent link styling)

3. **Add Quick Reference Elements**: Include "quick wins" or "minimum viable implementation" callouts for users seeking rapid compliance paths, reflecting the AI guidance's emphasis on practical implementation.

4. **Strengthen Accessibility Features**: Beyond eliminating hidden content, ensure:
   - Landmark navigation for screen readers
   - Skip navigation links
   - Consistent heading hierarchy
   - Alt text for any diagrams or visual elements

## Critical Success Factors for Phase 2

1. **Visual Distinction Execution**: Successfully implementing clear visual distinction between normative and guidance content without breaking integration benefits.

2. **Navigation Aid Implementation**: Delivering all promised navigation features (expandable TOC, breadcrumbs, cross-references) with consistent behavior.

3. **Accessible Markdown Formatting**: Ensuring the custom markdown style maintains accessibility while providing visual distinction.

4. **Content Completeness**: Successfully retrieving and integrating all relevant content from 30 source documents without gaps.

5. **Citation Preservation**: Maintaining all DocRef citations for traceability while integrating them naturally in the text flow.

## Final Validation Verdict

### Readiness for Phase 2
- **Status**: Ready to proceed
- **Confidence Level**: High
- **Rationale**: The proposed structure strongly aligns with AI guidance principles of transparency, accessibility, and user-centered design. While the AI guidance doesn't directly validate documentation structures, the analogous principles strongly support the proposed approach. The structure addresses all critical issues identified in Phase 1 analysis while respecting the constraint on core standards text.

### Key Validation Points
1. **Accessibility compliance** through elimination of hidden content and single document approach
2. **Process transparency** through conformance-centered organization making assessment requirements clear
3. **Integration best practice** reflecting AI guidance's own integrated presentation approach

### Implementation Guidance for Phase 2
- **Priority 1**: Implement clear visual distinction patterns before beginning content integration
- **Priority 2**: Build navigation structure early to test with sample content
- **Priority 3**: Validate accessibility at each stage rather than retrofitting
- **Priority 4**: Test progressive disclosure within sections during content creation
- **Priority 5**: Maintain citation traceability meticulously throughout content synthesis

## Phase 1 Completion Summary

Across 7 stages of analysis, we've identified that the identification standards' fundamental problem is misalignment between user goals (achieving conformance) and content organization (conceptual topics). The proposed single consolidated document with conformance-centered organization directly addresses this misalignment while implementing all eight major recommendations from our analysis. The structure is validated against AI guidance principles and ready for Phase 2 implementation.

## Next Steps

1. **Tom's approval checkpoint** - Review Phase 1 findings and structure proposal
2. **Approve or refine structure** based on Tom's feedback
3. **Proceed to Phase 2** - Begin with Stage 8 (Markdown Style Familiarization)
4. **Implement structure** with careful attention to critical success factors
5. **Validate at milestones** throughout Phase 2 content creation