Evidence Summarization

The AI Evidence Summarizer extracts key information from uploaded evidence files (PDFs, spreadsheets, documents) and generates concise summaries for workpaper narratives.

How It Works

1. Upload evidence to a workpaper or engagement
2. Click AI Summarize on the evidence item
3. The AI processes the document and generates:
   • A concise summary of the content
   • Key data points extracted
   • Relevant quotes or passages highlighted
4. Review the summary and click Apply to insert into the workpaper

Supported File Types

Format	Summarization Support
PDF	✅ Full text extraction + summarization
DOCX / DOC	✅ Full text extraction + summarization
XLSX / XLS	✅ Tabular data analysis + key metrics
CSV	✅ Data pattern detection
TXT	✅ Full text summarization
Images	⚠️ OCR-based (quality dependent)

AI Governance

• Summaries are always suggestions — never auto-inserted
• The original evidence file is never modified
• All AI interactions are logged with AIRequest/AIResult records
• Confidence scores are provided
• Citations reference specific pages or sections of the source document

Best Practices

• Use clear, text-based documents for best results
• Scanned PDFs may produce lower-quality summaries (OCR limitations)
• Always verify AI summaries against the source document
• Combine AI summaries with your own analysis

Permissions

Requires ai:suggest permission — available to Auditor, Manager, CAE roles.