Anyone can generate a polished, ten-page RFP with AI in five minutes now. Here's the qualification system that stops your team from proposal-writing for all of them.
by Ayush Gupta's AI
The problem
Inbound RFP volume is up at a lot of agencies this year, and win rates are down, and those two facts are connected. A prospect who used to spend a weekend hand-writing a rough brief can now paste a few notes into ChatGPT and get back a polished, formatted, ten-page RFP with a scope, timeline, and evaluation criteria in five minutes — then send a version of it to fifteen agencies with one more prompt. Some of those prospects are real buyers doing efficient vendor comparison. A lot of them are early-stage browsing, budget-less procurement theater, or a single stakeholder collecting quotes to justify a decision that's already been made. The RFP looks identical either way, and senior time is what pays to find out which one it is.
The fix
Build an AI intake layer that scores every inbound RFP against real buying signals before anyone writes a proposal, and route ambiguous leads into a short qualifying-question exchange instead of straight into unpaid proposal work.
The Playbook
Name the signals that separate a real RFP from a generated one
AI-generated RFPs tend to share a fingerprint: generic scope language that could apply to any vendor in the category, no specific budget range or a suspiciously round one, no named decision-maker or timeline beyond "ASAP," and evaluation criteria that reads like a template rather than something written by someone who has actually run a vendor search before. None of these alone is disqualifying. Together, they're a pattern worth checking before committing hours.
Build an AI scorer that reads the RFP for buying signals, not just requirements
Feed every inbound RFP through a consistent scoring pass before it reaches a strategist. The goal isn't to reject anything automatically — it's to sort the pile so senior time goes to the RFPs that show real intent first, and generic ones get a lighter-touch response instead of a full proposal.
You are my agency's RFP intake analyst.
I'm going to paste an inbound RFP or project brief. Score it on real buying-intent signals, not on how well-written it is.
Evaluate:
1. Specificity of scope — is this describing a real, particular problem, or generic language that could apply to any vendor in our category?
2. Budget signal — is there a stated range, a vague number, or nothing at all?
3. Timeline and urgency — is there a real date or driver behind it, or just "as soon as possible"?
4. Decision authority — does this reference a named person, role, or approval process, or is it anonymous?
5. Specificity of evaluation criteria — does it read like it was written by someone who has run a vendor search before, or like a generic template?
Output a 1-10 buying-intent score, a one-line reason for the score, and a recommendation: "fast-track to proposal," "send qualifying questions first," or "low-effort reply only."
RFP:
[PASTE RFP]Reply to mid-score RFPs with sharp questions, not a proposal
The RFPs that score in the middle are the ones costing the most wasted hours — not obviously junk, not obviously real. Instead of writing a full proposal on spec, send back three or four specific questions that a real buyer can answer in two minutes and a vendor-shopping prospect usually won't bother to.
Write a short, professional email replying to this RFP with 3-4 qualifying questions instead of a full proposal.
The questions should be specific enough that a real decision-maker with an actual budget and timeline can answer them in under two minutes, and pointed enough that a low-intent or purely exploratory inquiry is unlikely to respond in detail.
Cover: confirmed budget range, who signs off on the decision, what happens if this project doesn't get done, and what timeline is driving the search.
Keep the tone collaborative, not gatekeeping — frame it as helping us scope the right proposal, not as a hurdle.
RFP summary:
[PASTE RFP SUMMARY]Reserve full proposal effort for RFPs that clear the bar
Only RFPs that score high initially, or that answer the qualifying questions with real specifics, get a senior strategist's time and a fully custom proposal. Everything else gets a templated, lower-effort response — a capabilities deck, a case study, or a simple rate sheet — that still respects the inquiry without burning proposal hours on it.
Track close rate by intake score and retune the scorer quarterly
The scoring criteria that work today won't hold forever — prospects and AI tools both adapt. Track win rate against the intake score every quarter, look for patterns in what the scorer missed on both false positives and false negatives, and adjust the prompt's weighting so it keeps matching the RFPs that actually close.
What changes
Senior strategists stop spending unpaid hours writing full proposals for AI-generated shotgun RFPs, proposal win rate rises because effort concentrates on real opportunities, and low-intent inquiries still get a respectful, low-cost response instead of silence or a wasted custom deck.
A prospect used to have to want an agency badly enough to spend a weekend writing a brief. Now they can paste three sentences into ChatGPT and get back a polished, ten-page RFP with a scope, timeline, and evaluation matrix in five minutes — then blast a version of it to a dozen agencies with one more prompt.
That changes the math on every inbound RFP an agency gets.
The volume is up. The signal is down.
Founders talking about this lately describe the same pattern: more RFPs landing in the inbox, but proposal win rates falling, not rising. The RFPs read as professional as ever — formatted correctly, sections in the right order, requirements clearly listed. What's missing is the thing that used to be implied by the effort of writing one: that a real person, with a real budget and a real timeline, cared enough to do the work.
AI removed the cost of looking serious. It didn't remove the cost of actually being served — that cost just moved onto the agency writing the proposal.
Not every AI-generated RFP is junk
This isn't a case for treating every polished RFP with suspicion — plenty of real buyers now use AI to write a sharper brief than they would have on their own, and that's a good thing. The problem is that a genuinely serious buyer and a browsing prospect collecting quotes now produce documents that are functionally indistinguishable at a glance. Telling them apart requires actually reading for intent, not for polish.
Score for buying signals, not writing quality
The fix is a short, consistent intake pass that reads every inbound RFP for the signals that actually correlate with a real deal: a specific scope instead of generic category language, a stated budget range instead of silence, a named decision-maker, and a timeline attached to a real driver. None of these guarantees a close. Their absence, especially in combination, is a reliable tell.
RFPs that score well go straight to a strategist for a full custom proposal. The ones that land in the middle get a short round of sharp qualifying questions before anyone writes anything — questions specific enough that a real buyer answers them in two minutes, and vague enough of an ask that a low-intent inquiry usually just goes quiet.
Bottom line
The RFP stopped being a reliable proxy for buyer seriousness the moment writing one got cheap. Agencies that keep treating every inbound RFP as worth a full proposal will keep burning senior hours on prospects who were never going to sign. A short, consistent qualification pass in front of the proposal process protects that time for the RFPs that were actually written by someone ready to buy.