rfp-responder

name: rfp-responder description: Use when a rep receives a formal RFP, RFI, or security questionnaire. Parses the document into a requirement tree, maps each item to our capabilities, flags must-have gaps early, and drafts compliant section-by-section responses using the customer's language and cited evidence. safety: writes-local

When to use

Trigger this skill when:

• A buyer submits a formal RFP, RFI, or RFQ document (Word, PDF, or spreadsheet).
• A procurement team sends a security questionnaire (SIG, CAIQ, or custom).
• A due-diligence buyer (private equity, M&A target) sends a request for information.
• An existing customer's procurement triggers a vendor review with a questionnaire.

Do not use this for informal buyer emails asking general questions — use a targeted email reply instead. The RFP responder is for structured, multi-section documents.

Inputs

Required:

• rfp_document — the RFP text, file path, or pasted contents. Must be complete.
• our_capability_library — pointer to our answer bank, case studies, security docs. Typically in Notion or a dedicated RFP tool.
• customer_context — who is issuing the RFP, their stated objective, their evaluation timeline.

Optional:

• prior_responses — previous responses to similar RFPs, as raw material.
• deal_owner_guidance — rep's notes on what to emphasize, what to avoid.
• win_themes — strategic themes to thread through every section (e.g. "time to value," "vendor consolidation").
• submission_format — the format the buyer expects back (Word with their template, Excel with their rows, PDF).

Outputs

Three artifacts:

1. Requirement tree — structured extraction of every numbered/lettered requirement in the RFP, flagged as Must / Should / Nice / Informational.
2. Gap analysis — requirements we do not fully meet, categorized as Partial / Missing / Custom-buildable, with a recommended response strategy for each.
3. Draft response — section-by-section draft in the buyer's template, using the buyer's section numbering and language, with evidence cited per answer.

Tool dependencies

• Notion MCP or a dedicated RFP tool (Loopio, Responsive) — highly recommended. The answer bank is the skill's primary source.
• WebFetch — optional; pulls the buyer's public site for context.
• Salesforce / HubSpot MCP — optional; reads prior opportunities with the same buyer.
• Docx / PDF reader — required if the RFP is supplied as a file. If unavailable, the user must paste the content.

Procedure

1. Ingest the RFP. Extract every requirement. Preserve the buyer's numbering (2.3.1, §4(a)(ii), etc.) exactly. Do not renumber.
2. Classify each requirement.
   • Must — explicit must-have, pass/fail criteria.
   • Should — strongly preferred.
   • Nice — bonus.
   • Informational — disclosure only, no scoring. Classification is often stated directly ("mandatory requirement"); if implicit, flag for the rep to confirm.
3. Map each requirement to our capability. For each requirement, one of:
   • Full — we meet it natively.
   • Partial — we meet it with caveats.
   • Roadmap — planned within 6 months.
   • Custom — we can meet it via services/configuration.
   • Gap — we do not meet it.
4. Produce the gap analysis early. Before drafting any prose, surface Must + Gap combinations as deal-threatening items. The rep needs 24-48 hours to make the go/no-go call on these.
5. Draft responses section by section.
   • Use the buyer's section headings, exactly.
   • Use the buyer's language — if they say "data lineage," don't respond about "provenance."
   • Lead each answer with a direct yes/no/partial where applicable.
   • Follow with one paragraph of explanation and one citation of evidence.
6. Cite evidence inline. Every substantive claim must reference one of:
   • A named case study.
   • A security document (SOC 2 report, ISO 27001 cert, Pen Test summary).
   • A feature documentation page or API reference.
   • An architecture diagram. Citations are explicit: "Evidence: See Case Study — Acme Retail (Appendix B, page 12)."
7. Thread win themes. If provided, weave them through the executive summary and into at least three answer bodies. Do not force them where they do not fit.
8. Write the executive summary last. It references the themes that emerged from the answers, not the other way around.
9. Handle security / compliance sections with precision. Answer exactly what is asked. Do not volunteer more. Over-disclosure creates procurement friction.
10. Produce the output in the buyer's submission format. If they supplied a Word template with specific fields, output those fields filled. Do not rewrite the document structure.

Examples

Example 1 — RFP response for Nimbus Logistics visibility platform

Inputs: 47-page RFP from Nimbus Logistics covering functional, non-functional, security, and commercial requirements. 142 numbered requirements total. Customer context: new CIO consolidating visibility spend, evaluation deadline 4 weeks.

Output structure:

Gap Analysis (delivered first, Day 1)

• Must-haves we fully meet: 118 of 142.
• Must-haves we partially meet: 19. Recommended strategy per item listed below.
• Must-haves we do not meet: 5. These are deal-threatening.
  • §3.4.2: "Real-time vehicle telemetry ingestion at 100Hz per asset." — Our current platform supports 10Hz. Recommended response: disclose gap, propose roadmap commitment, price the custom work separately.
  • §5.1.7: "On-premises deployment option with no cloud dependency." — We are cloud-native. Recommended response: propose dedicated VPC deployment as alternative. Flag for CRO go/no-go.
  • [...]

Draft Response (excerpt)

§3.1.1 — Requirement: "The platform shall provide real-time visibility into in-transit shipments across ocean, rail, and road modalities."

Response: Yes. Our platform ingests carrier and tracking data across all three modalities, with a median event latency of 45 seconds and a 99th percentile of 3 minutes. Evidence: Architecture overview, Appendix C, page 4. Reference customer: Maersk — 2024 case study (Appendix D).

Example 2 — Security questionnaire for a mid-market SaaS buyer

Inputs: 180-item CAIQ-style questionnaire from Helios Analytics. Output is a spreadsheet-shaped response with Y/N/Partial columns, explanation column, and evidence column. Skill flags that 4 items require net-new attestation from our CISO's team before submission. Draft responses for the remaining 176 are produced.

Constraints

• Preserve the buyer's numbering exactly. Renumbering is a procurement red flag.
• Use the buyer's language. If they say "supplier," do not reply "vendor."
• Every substantive claim has an evidence citation. No unsupported assertions.
• Must-have gaps are surfaced in the gap analysis on Day 1, before prose drafting.
• Do not over-disclose on security questions. Answer what is asked.
• If the answer is "no," say so. Buyers detect evasion and score it worse than an honest gap.
• No marketing superlatives in RFP prose. "Industry-leading" and "best-in-class" are banned.
• The executive summary is under 500 words.
• Every answer lists the version of evidence cited (e.g. SOC 2 Type II report, dated YYYY-MM-DD).

Quality checks

Before returning:

• Every numbered requirement from the RFP appears in the requirement tree.
• Every requirement is classified Must/Should/Nice/Informational.
• Every requirement is mapped Full/Partial/Roadmap/Custom/Gap.
• Gap analysis lists all Must + Gap combinations as deal-threatening.
• Draft uses the buyer's headings and numbering.
• Draft uses the buyer's vocabulary — verified against a quick diff of terms they used vs. terms we used.
• Every answer starts with a direct yes/no/partial.
• Every answer cites specific evidence.
• Security section answers only what is asked.
• Output is in the buyer's requested format.
• Win themes appear in the executive summary and at least three answer bodies (if provided).