Free skill pack
Stop writing proposals from a blank page.
A tool-agnostic recipe for founders and owners who still write every proposal themselves. Build a library of reusable blocks and a plain-language pricing rules document once, then assemble each new proposal with one saved prompt in ChatGPT, Claude, or Gemini. The draft arrives with every guess flagged, so your job shifts from writing to reviewing and pricing.
Free download · Markdown recipe
Proposal draft workflow
A markdown recipe you can run today: prerequisites, a one-time setup that turns your best past proposals into a block library, the full assembly prompt, and a review checklist that catches invented prices and scope before the client sees them.
- Download the recipe and open it in any text editor.
- Run the one-time setup: extract a block library from past proposals and write your pricing rules.
- For each new deal, paste both documents plus your client notes into your AI assistant with the assembly prompt.
- Work through the review checklist, resolve every flag, then send.
What this handles
The recipe takes three inputs: a block library built from your best past proposals, a pricing rules document written in plain language, and your notes on the deal at hand. It produces a first-draft proposal plus a "Review before sending" list of every gap it found.
The constraint doing the real work: the model is forbidden to invent anything. It can only fill placeholders from your client notes or your pricing rules. Anything unsourced becomes a [MISSING] flag, and any deal that doesn't fit a pricing rule becomes a [PRICING: needs manual quote] flag instead of a made-up number.
Your blocks stay in your wording. The prompts explicitly tell the model not to improve the language that already won you work, only to adapt the transitions between blocks.
How to run it
Setup happens once. You paste two or three past proposals into the assistant with the extraction prompt, hand-edit the block library it returns, and write your pricing rules one per line: what you charge, how it scales, what triggers a custom quote, what you never discount.
The per-deal loop is short. Fresh chat, paste the block library, the pricing rules, and the client notes, then run the assembly prompt. Read the "Review before sending" list first and hold one rule without exception: the model assembles, you price.
The review checklist in the download covers the rest: resolve every flag, check each scope sentence against what the client actually asked for, and confirm no name or deliverable leaked in from the proposals your library was built from. If the same flag appears deal after deal, fix the library or the rules document, not the prompt.
When to upgrade
This is a manual loop, and that is fine while one person owns proposals and volume is low. It stops being fine when several people run quietly diverging copies of the library, when quotes depend on live data from your systems like rates, inventory, or availability, or when one mispriced proposal costs real money. At that point the fix is not a better prompt. It is a system that drafts from the CRM record, prices from your source of truth, and routes every proposal for approval before it leaves. That is a build, not a snippet. See what we install or book a free AI opportunity audit to find out whether your proposal volume justifies it.
Want proposals drafted before you open the doc?
We install agents that draft from the CRM record, price from your source of truth, and route every proposal for approval before it leaves. Same team. Double the output.
We take on companies ready to invest $5,000+/month. Not there yet? Our free resources are genuinely free.