The End of the RFP and the Rise of Scope — Modern Tactics for the Money Fight
Introduction
People are in place. Technology decisions are made. Now comes the final reckoning: money — the cold reality of budgets and scope.
Twenty years ago, the number-one fuse for project explosions was the "vague RFP (Request for Proposal)." A client would hand over a few pages saying "build something like this," and the vendor would fill in the gaps with assumptions and produce an estimate. The misalignment began at that very moment.
Remarkably, the fundamental dynamic hasn't changed in twenty years. But the tactics on the ground have evolved significantly.
This article distills the essence of Chapters 4 and 5 from the Zenn series "Ultra-Practical! IT Project Management That Doesn't Fail — Remake" into a modern, actionable framework.
The 20-Year Curse of Lump-Sum Estimates
When organizations commit investments worth millions or tens of millions of dollars, "we won't know until we try" doesn't fly. The result: even today, we're trapped in the structural contradiction of having to produce "precise estimates" for an uncertain future — effectively telling a well-dressed lie.
Large-scale projects, in particular, still demand "submit a proposal and estimate based on a detailed RFP." Corporate governance and budget planning processes require it. This remains an immovable reality.
The RFP continues to function as a powerful ritual.
Three Tactics for Bridging the Gap Between Contract and Reality
Twenty years ago, "build exactly what the RFP says" was the correct answer. Today, the assumption is that "what was written in the RFP will become outdated during development."
The modern PM's battle is a high-wire act: achieving adaptive development within the rigid shell of fixed-price, lump-sum contracts.
1. Phased Contracts (Carving Out Risk)
Bundling "requirements and design" with "development and implementation" into a single contract is a recipe for disaster. Carve out the high-uncertainty front half as a time-and-materials engagement, then negotiate the second half once requirements are clearer. This single step dramatically improves estimation accuracy and risk control.
2. Strategic Use of Buffers (Agreeing on Margins)
Building buffers into estimates is not "dishonest." In an era where change is the default, buffers are an essential cost for maintaining quality and adapting to change. The PM's real skill lies in transparently explaining this to the client and securing their agreement.
3. Granular Acceptance (Cash Flow Defense)
Waterfall-style single-point acceptance means that delivery delays can be fatal for vendors — delayed payments lead to cash flow crises. Build cycles where working deliverables are reviewed and accepted as frequently as possible. This enables course corrections even when the RFP becomes outdated, preventing the final catastrophic blowup.
It's especially important for the client side to establish stage-gate investment decisions — "if the PoC phase doesn't produce results, we don't proceed to full development." Not a kamikaze charge that can't be stopped once started, but the courage to calmly evaluate at each phase whether to continue or cut losses. That's the cornerstone of modern risk management.
LLM-Generated RFPs: A New Trap
LLMs have dramatically accelerated RFP prototyping. What used to take weeks to concretize can now be done in hours.
But a new trap has emerged: the "LLM-generated RFP."
Now that clients are also using AI to write RFPs, documents that look remarkably polished and precise are appearing. But closer inspection reveals logical inconsistencies, or unrealistic SLAs and security requirements quietly embedded — all wildly out of proportion to the actual budget.
Because these RFPs "look perfect," they're in some ways worse than the flimsy RFPs of twenty years ago. Accept them at face value and your estimate will be wildly off-budget. Force-fit it to the budget and the project is guaranteed to die.
What's now required of PMs is defensive yet substantive scope management: "See through the plausible lies in AI-generated RFPs, use AI to prove value quickly within a fixed budget, and reprioritize investments accordingly."
Not Sacrifice, but Investment — Understanding MVP Correctly
Twenty years ago, when a project hit crisis, the PM's job was to coldly choose "what to give up." Cut features? Extend the deadline? Squeeze out more budget? It was a "sacrifice" — a painful, negative decision.
Fast forward twenty years. In a world where uncertainty is the norm, that framing has shifted. "Sacrificing" something has been elevated to an everyday "choice."
Trying to Include Everything Means Aiming at Nothing
"We're cutting this feature" versus "We're going all-in on making this experience exceptional, so we're deferring the rest for now." This isn't just wordplay — it represents a fundamental evolution in product management.
MVP Is Not "Cutting Corners"
An MVP is not merely an "incomplete product." It should be a sharpened blade, honed to prove the hypothesis you believe in. The decision to strip away unnecessary features is the most creative act of all — it forces clarity on what you want users to see.
The Responsibility of Deciding
AI can give you the "average right answer," but it won't take responsibility for close calls. Sifting through mountains of data, then trusting your intuition and conviction to point and say "this way" — that resolve is what moves teams.
Summary
Twenty years ago, an RFP was "a letter." You mailed it and waited for a reply.
Today, an RFP is merely "the start of a conversation."
The shared goal should be "how do we solve the user's problem?" rather than "which buttons go where." Instead of piling up detailed requirements documents, maintain a daily dialogue in GitHub Issues or Notion about "what's valuable right now."
The fight over money is no longer waged on contracts — it's waged on team trust.
Translating the lessons found through blood and sweat on the front lines twenty years ago into today's context:
- Avoid fatal wounds (Resilience) — Protect yourself with phased contracts and buffers
- Fail fast and learn (Fail Fast) — Run short cycles through granular acceptance
- Bet on what truly matters (Focus) — Sharpen your hypothesis with MVP
This is project management for an era that assumes the challenge.