The complete RFP response process,
in six stages.

You don’t lose RFPs because you can’t write. You lose them because by the time you’re writing, half the inputs are wrong, the scope is unclear, and your team hasn’t weighed in.

This guide is the antidote: a six-stage process that pushes 80% of the work into preparation, leaving the writing as the comparatively easy part. It’s the same architecture we’ve baked into qlows, but it works whether you use a tool or run it on whiteboards.

When teams describe a great bid, they describe its prose. When teams describe a fastgreat bid, they describe the preparation. There’s a reason for that.

• Prose is high-leverage but only when the inputs are right.
• Inputs are slow to gather but compound across bids — every answered question is a library entry for next time.
• AI accelerates writing dramatically but is useless without grounded context.

So the discipline is: spend disproportionately more time on the first four stages (Triage → Intelligence → Compliance → Field) than on the actual drafting. The drafting then takes hours, not days.

If you’re writing the bid in the last 48 hours, you didn’t do enough preparation. If you’re writing the bid in the first 48 hours, you didn’t do enough preparation either — you’re hallucinating the inputs.

The single most important hour of the RFP cycle. It decides whether you should be in this race at all.

Things to look for in a 60-minute triage:

• Wired-RFP signals.Incumbent re-bids, hyper-specific tech requirements, copy-pasted clauses from a competitor’s website. If 3+ signals fire, the bid is probably allocated and you’re cover.
• Mandatory disqualifiers.Geo restrictions, certifications you don’t hold, references in markets you’ve never served. One missing mandatory = automatic disqualification.
• Margin.Hard cap on price? Mandatory T&M? Run a 15-minute back-of-envelope before committing.
• Strategic fit. Does this win unlock the next three bids? Is this a logo you actually want?

If you walk away here, you’ve saved 40 hours. If you go, go decisively — full team mobilization, not half-hearted drafting.

Read the RFP end to end. Then read the appendices. Then read the buyer’s last public RFP (if available) to understand their style.

What you’re extracting:

• Compliance items— every “the vendor must” / “the proposal shall” sentence becomes a row.
• Evaluation criteria — explicit weighting (often 30% technical / 30% experience / 25% pricing / 15% references — varies wildly).
• Risk signals — penalty clauses, non-standard termination, unusual liability caps.
• Format requirements — page limits, font, section order, file naming conventions. Boring, but disqualifying.

A 100-page RFP yields ~50 compliance items, ~10 evaluation criteria, and 2-5 risk flags. Get all of them on paper before stage 3.

The compliance grid is the artifact that earns the most ROI per minute spent. It’s a table:

• One row per requirement — verbatim from the RFP.
• Section reference — where in the RFP it lives.
• Owner — which team member can answer.
• Status — open, drafted, reviewed, locked.
• Response location — which section of your response document addresses it.

This grid is the source of truth for “are we done?”. When every row is locked and has a response location, you’re compliant. Until then, you’re not — and the bid manager should know which rows are blocked on whom.

This is where bids actually get won or lost. Three columns:

• Client— what does this buyer actually care about? What’s the political subtext? Who will read this?
• Competitors— who else is bidding (you can usually guess from the RFP’s style)? What’s their obvious pitch? Where are they weak?
• Positioning — your three differentiators, ranked, that this buyer specifically will value.

If you can’t articulate your top differentiator in one sentence here, the bid won’t articulate it either. Don’t move past stage 4 until the field is locked.

By now you have a compliance grid with rows you can’t answer alone. Maybe the CFO needs to confirm pricing tier; Legal needs to weigh in on liability; the Tech Lead has to confirm a specific integration capability.

The trap: assume those people will check Slack and respond. They won’t — they’re busy with their day job. So route questions explicitly, with deadlines, in a surface they can answer in one click.

What “route” actually means:

• One question (or a small batch) sent to one person via a secure link.
• They see only their questions — not the whole bid.
• They answer in their own surface, in their own words.
• The answer is captured at the source, attributed, time- stamped, and pulled into the response automatically.

See the collaboration guide for the deep dive on this.

Now — and only now — does the writing start. The compliance grid is locked. The field is locked. The team has answered. You’re holding a complete prep pack.

The AI (Claude, Cursor, ChatGPT, whatever you use) gets handed this prep pack as context. The drafting becomes a 3-4 hour exercise instead of a 14-hour one — because the AI has every input it needs, grounded in your past proposals and verified team answers.

A small but real point: the AI doesn’t need to be ours. We’re MCP-native because we believe the writing tool should be your choice. qlows’ job is to make the prep pack so good that whatever AI you use produces a tight, grounded first draft.

Reserve the last 24 hours for nothing but quality control:

• Compliance grid → response document cross-check. Every row locked? Every reference accurate?
• Format pass — page count, font, file naming, signatures.
• One human read-through, ideally by someone who didn’t write it.
• Submit at least 4 hours early. Portals fail. Email queues back up. Don’t cut it close.

Win or lose, do a 30-minute post-mortem. The data you capture here is the input library for the next bid:

• Which compliance items were hardest to answer? (Library entries for future RFPs.)
• Which team members were the bottleneck? Why?
• Which differentiators landed? Ask the buyer if you lost.
• How accurate was the time estimate? What stage took longer than expected?

Done well, every bid leaves you 30% faster on the next one. That’s the compounding curve.

The whole architecture above is a process, not a tool. You can run it on Notion + Slack + Excel if you have the discipline. qlows exists because that discipline is brutally hard to maintain across 4-15 RFPs a month — and the tools you already use weren’t designed for this work.