Skip to content

Insurance Form Question Extraction Summary

Extraction Date: 2025-06-13 (Updated) Total Questions Extracted: 127 (was 113)

Summary by Insurance Company

Company Forms Total Questions
AIG 1 35
Blue Zebra 1 19
Chubb 3 46 (+14)
Liberty 1 6
Unknown (Combined) 1 21

Question Categories by Risk Pattern

Risk Pattern Count Percentage
general 91 71.7%
incident 9 7.1%
governance 8 6.3%
employee 7 5.5%
third_party 7 5.5%
technical 6 4.7%
data_breach 2 1.6%
business_interruption 1 0.8%
ransomware 1 0.8%

Question Types Distribution

Answer Type Count Percentage
text_short 80 63.0%
numeric 15 11.8%
yes_no 12 9.4%
date 10 7.9%
multi_choice 5 3.9%
text_long 4 3.1%
dollar_amount 1 0.8%

Framework Mapping

Framework Count Percentage
general 117 92.1%
essential_eight 4 3.1%
nist 3 2.4%
privacy_act 2 1.6%
iso27001 1 0.8%

Key Observations

  1. Question Quality: Some questions were fragmented due to PDF extraction challenges. Manual review recommended for:

  2. Questions that appear incomplete (e.g., "are documented.")

  3. Questions missing their beginning (e.g., starting mid-sentence)

  4. Coverage Areas: The forms heavily focus on:

  5. General compliance questions (72%)

  6. Technical security controls
  7. Incident response capabilities
  8. Governance and policy questions
  9. Employee and third-party management

  10. Answer Types: Majority require short text responses (63%), with numeric (12%) and yes/no (9%) following.

  11. Framework Alignment: Most questions are general compliance (92%) rather than specific framework-aligned, suggesting opportunity for framework mapping enrichment.

New Chubb Form Analysis

The Chubb Forefront Portfolio form added 14 questions focusing on:

  • Employment practices liability
  • Crime controls and social engineering fraud
  • Privacy compliance (Australian Privacy Principles)
  • Backup and security fundamentals

Files Generated

  1. insurance-questions.json - Full structured data for import
  2. insurance-questions.csv - Spreadsheet format for review
  3. insurance-questions.md - Human-readable format with all questions

Next Steps

  1. Manual review to fix fragmented questions
  2. Enhance framework mapping with more specific control references
  3. Add severity/criticality scoring to questions
  4. Cross-reference with existing question bank to identify duplicates
  5. Create category-based question sets for different business types

Notes for Import

  • All questions have been tagged with the 5 required attributes as specified in outstanding-work-tracker.md
  • Risk patterns and framework mappings are stored as arrays to support multiple values
  • Source file and extraction date included for traceability
  • Page numbers preserved for reference back to original documents