AI Narrative Generation

The AI narrative engine generates professional audit report content from structured data. It transforms findings, test results, and engagement metrics into clear, well-structured prose.

Capabilities

Feature	Input	Output
Executive Summary	Engagement outcomes, findings, metrics	2-3 paragraph executive overview
Introduction	Engagement objective and scope	Professional introductory narrative
Findings Narrative	CCCER finding data	Structured finding descriptions
Conclusion	Overall results, rating	Concluding assessment
Section Drafting	Section topic + underlying data	Custom section content

How to Use

1. Open a report in edit mode
2. Navigate to the section you want to draft
3. Click the AI Draft button (✨)
4. Review the generated text in the preview panel
5. Edit as needed
6. Click Apply to insert into the report

Data Sources

The AI uses the following data to generate narratives:

• Engagement metadata (objective, scope, timeline)
• Linked findings with CCCER components
• Test procedure results (pass/fail counts)
• Evidence summaries
• Action plan status
• Historical engagement data (via RAG)

Quality Controls

• AI-generated text always appears as a suggestion — never auto-inserted
• Confidence score displayed with each suggestion
• Citations reference specific findings, workpapers, or evidence
• Human must explicitly Apply or Reject each suggestion
• All interactions logged with AIRequest and AIResult records

Best Practices

Do	Don't
✅ Use AI as a starting point, then personalize	❌ Submit AI text without review
✅ Check that all findings are accurately represented	❌ Rely on AI for severity judgments
✅ Verify citations reference the correct workpapers	❌ Skip the citation check
✅ Maintain professional audit language	❌ Leave informal AI phrasing