Brief sections
The leak
This one isn’t on our twelve-leak list; it earned its own brief because it keeps showing up in the same businesses. An RFP lands with a two-week window. Days one through nine, nothing happens — everyone’s busy and starting feels heavy. Days ten through fourteen are a scramble: copying from the last proposal, hunting for the paragraph that describes your safety program, rebuilding a compliance matrix at midnight. Some bids don’t go out at all — the silent no-bid, where the pipeline shrinks not by decision but by exhaustion. And the founder is the bottleneck, because the proposal voice lives in their head.
What AI realistically does here
Assembly, coverage, and first drafts — not persuasion. The system ingests your past proposals, your SOP library, and the RFP itself; builds the requirements matrix so nothing gets missed on a technicality; and drafts the standard sections in your established voice, with your actual capabilities as source material. The senior person’s job compresses to the part that wins or loses the work: strategy, pricing, and the two paragraphs that are genuinely new. What it can’t do: decide whether to bid, invent differentiation you don’t have, or write the client-specific insight that makes a proposal land. A proposal that’s entirely generated reads entirely generated, and evaluators can tell — the value is starting from mostly-done, not shipping unread output.
What has to be true first
- A corpus: past proposals, ideally with won/lost noted even roughly. Your old proposals are the source material for your voice and your claims — no corpus, no shortcut.
- Claims discipline: the boilerplate has to be true. Drafting from your SOP library keeps capability claims tied to how you actually work, because automation amplifies whatever’s in the source material, including the exaggerations.
- One owner of the proposal voice with final-read authority. Faster drafts with no editor just means faster mediocrity.
What we'd install
Proposal drafting gets scored in the opportunity portfolio against your other leaks — it ranks high where bid volume is real and the silent no-bid is quietly costing pipeline. The proposal library and a claims register go into the SOP library, the drafting copilot joins the roster owned by whoever owns business development, and the value ledger tracks drafting hours per proposal plus the number owners feel most: bids submitted that would previously have been skipped. Sequence on the AI operating partnership.
Worth it?
Count the senior hours in your last three proposals, and the bids you didn’t submit last quarter because nobody had the week to spare. The calculator below prices the hours as an illustrative annual figure; the skipped bids are a pipeline number only you can estimate, and it’s usually the larger one. The scorecard places this against your other leaks — if quoting is also broken, that usually comes first, since proposals inherit whatever your pricing process produces.