Security Questionnaire turns an XLSX buyer questionnaire and an explicitly approved evidence set into cited answer drafts, visible unknowns, and a reviewed workbook. It is built for high-stakes enterprise sales work where a plausible but unsupported answer is worse than no answer.
The workflow preserves original spreadsheet structure, binds each draft to source IDs, and makes conflicts and unsupported questions visible. It never infers certifications, audit scope, product coverage, data location, subprocessors, customer identity, or who has authority to approve an answer.
What you can do
- build_answer_library — turn approved policy text into traceable evidence chunks
- analyze_questionnaire — detect questions and answer cells while leaving the workbook untouched
- draft_answers — draft only from selected evidence and return citations, confidence, and review states
- review_conflicts — separate drafts from conflicts, unknowns, and answers needing review
- export_questionnaire — write only explicitly human-approved answers into a copied XLSX
Who it's for
B2B SaaS security teams, sales engineers, GRC leaders, and founders responding to customer due-diligence questionnaires without weakening the evidence trail.
How to use it
- Extract policy text with the document tools and mark only approved sources
- Analyze the original workbook and confirm the detected question and answer columns
- Draft answers, then resolve every conflict and unknown with the responsible evidence owner
- Have a designated human approve individual answers
- Export a copied workbook; send or upload it through your normal reviewed process
Getting started
Call analyze_questionnaire on a private signed XLSX URL, then pass its questions and approved source text to draft_answers. The tool creates drafts only: it has no portal-login or external-submission capability.
