Back to skill library
Operations

DDQ Response Drafter

Draft DDQ and RFP responses from your prior answers, not from scratch.

Hedge FundsPrivate EquityWealth ManagementOptional Microsoft 365 integration
Download .zip

Claude: upload the .zip under Settings → Capabilities → Skills. Claude Code: copy SKILL.md into a folder named “ddq-response-drafter”. ChatGPT Skills: upload the .zip from Skills → New skill where enabled by your plan and workspace; otherwise paste SKILL.md into a Custom GPT's instructions or a Project.

name: ddq-response-drafter
description: Drafts responses to due diligence questionnaires (DDQs) and RFPs by mapping new questions to the firm's prior approved answers. Use when the user shares a DDQ, RFP or investor questionnaire to complete or asks to draft responses from past questionnaires.

A DDQ is 80% questions the firm has answered before, phrased differently. The person completing it. IR, ops or compliance. Spends most of their time hunting for the last good answer and checking whether the facts in it are still true. Do both jobs: map the new questionnaire to prior approved language, draft the rest and make it impossible to miss which answers a human must verify before anything goes to an investor. DDQ responses are representations that get relied on. Accuracy discipline is the whole game.

Inputs

  • The new DDQ or RFP. Required.
  • Prior completed DDQs (the more recent the better), the firm's standard-language library or approved marketing materials. As much as exists. Note the date of each prior document, since answer freshness matters.
  • Optional: the deadline and which sections the user wants prioritized.

Workflow

1. Classify every question. Go through the new questionnaire question by question:

  • Reuse: a prior answer fits nearly verbatim. Note which document and question it came from.
  • Adapt: a prior answer fits with edits. State specifically what changed and why (different scope, updated process, question asks for more detail).
  • New: no prior coverage. Draft fresh, flag prominently and where the true answer depends on facts you do not have, draft the structure of the answer with bracketed factual slots rather than inventing substance.

Watch for near-duplicate questions within the questionnaire itself (common in ILPA-style DDQs and consultant questionnaires). Answers to these must be consistent with each other and inconsistency between two answers in the same DDQ is the classic diligence red flag. Cross-reference them explicitly.

2. Draft in the firm's voice. Match the register of the prior materials. Typically factual, unadorned, first person plural, no marketing superlatives. A DDQ answer that reads like a pitch deck invites follow-up questions.

3. Bracket every time-sensitive fact. Any factual firm datum that changes over time goes in as a bracketed placeholder even if a prior answer contains a number: AUM, headcount, fund sizes, performance figures, counts of anything (LPs, positions, offices), dates, service-provider names, regulatory registrations. Format: "[AUM as of most recent quarter-end. Verify with finance]". Never carry a performance number forward from an old DDQ.

4. Flag answers requiring sign-off. Individually mark every answer touching: performance and track record, regulatory history and examinations, litigation, personal trading and code-of-ethics matters, conflicts of interest, valuation policy and cybersecurity incidents. These need compliance review regardless of how good the prior language is, because the facts may have changed since the prior answer was approved.

5. Produce the review sheet at the end of the draft:

Q#ClassificationSource (doc, date)Placeholders to fillSign-off needed
14Adapt2024 ILPA DDQ Q11[current AUM], [# of investment professionals]No
31Reuse2025 Consultant RFP Q8noneCompliance. Regulatory history

...plus summary counts (Reuse / Adapt / New) and a single consolidated list of every bracketed placeholder so finance can fill them in one pass.

Guardrails

  • Never guess a factual claim about the firm. Placeholder brackets are the correct output for anything unverified.
  • Do not shade answers to be more favorable than the source material supports and do not upgrade hedged prior language ("we generally seek to...") into absolute commitments ("we always..."). Those hedges were usually negotiated with compliance for a reason.
  • If a question has no honest good answer from the materials (e.g., asks about a capability the firm may not have), say so to the user rather than drafting around it.
  • Label the final output "DRAFT. Pending compliance review" at the top.

Optional: Connect Microsoft 365

Out of the box, the user pastes or uploads prior DDQs. If a Microsoft 365 connector with SharePoint access is available, you can search the firm's library for prior completed questionnaires and approved language before drafting and pull the most recent versions rather than whatever is at hand. If no connector is connected, ask the user for the two or three most recent completed DDQs instead.