7 min read
How long does it usually take you to send a proposal after a good discovery call? If the honest answer is two or three days, and sometimes never because the moment passes and you lose the nerve, you are leaving work on the table that was practically yours. The client was warm. They asked for a proposal. And then the busy week swallowed it. This guide walks you through a workflow that turns that call into a finished, sent proposal in about thirty minutes, using AI for the heavy lifting and you for the parts that actually matter.
You do not need to be technical to set this up, and you do not need expensive software. You need one AI assistant, one place to send and sign the document, and a little preparation done once. Let me walk you through it step by step, the way I would if we were sitting together.
Who this is for and what it solves
This is for freelancers, consultants, and service providers, the coach, the designer, the fractional marketer, anyone who wins work by proposing it. The problem it solves is not writing quality, it is speed and consistency. The proposal that arrives within an hour of the call, while the client still feels the excitement of the conversation, closes at a far higher rate than the polished one that shows up next week. Speed is the edge here, and AI is very good at giving you speed without making you sound like a robot, as long as you keep your hands on the wheel.
What you will need before you start
Gather three things, and it will take you fifteen minutes to set up the first time and seconds every time after.
- An AI assistant you already pay for. ChatGPT, Claude, or Gemini all work well for this. Claude in particular is often praised for longer, reasoned documents like proposals and strategy briefs. If you have not settled on one, our guide to which twenty-dollar AI to actually pay for will help you choose.
- A place to send and sign. A dedicated proposal tool like Bonsai handles proposals, contracts with electronic signatures, and invoicing in one place, which is ideal for freelancers. A well-formatted document and an e-signature link also work fine to start.
- Your notes from the call. A recording and transcript is best, but five honest bullet points scribbled the moment you hang up are plenty.
The seven steps, start to finish
Step one: capture the call while it is fresh. The single biggest reason proposals go wrong is that we forget what the client actually said and fill the gaps with guesses. Use a meeting-notes tool to transcribe the call, or take ninety seconds right after to write down their exact goal, their deadline, their budget signals, and any specific phrases they used. Those phrases are gold, because echoing a client’s own words back to them is what makes a proposal feel personal. This is the same discipline that powers a good client onboarding flow: capture once, reuse everywhere.
Step two: build your proposal skeleton, one time only. Before you touch AI, write down the sections every one of your proposals should have. A common shape is: the problem in the client’s words, your recommended approach, what is included and what is not, timeline, three pricing options, and next steps. Save this as a template. You are building the mold once so you never start from a blank page again.
Step three: hand the AI your notes and your skeleton together. Open your assistant and give it a clear instruction. A prompt that works well looks like this: “You are helping me draft a client proposal. Here are my notes from the discovery call, and here is my standard proposal structure. Draft the proposal using my structure, use the client’s own words for the problem section, keep the tone warm and direct, and mark anything you are unsure about in brackets so I can fill it in. Do not invent scope or prices I did not give you.” That last sentence matters, and we will come back to it.
Step four: give it the pricing, do not ask it for the pricing. Tell the AI your three tiers and what each includes, and let it format them cleanly into a good, better, best layout. Offering three options rather than one number reliably raises the average deal size, because the client stops deciding whether to hire you and starts deciding how much. The AI is excellent at structuring and describing the tiers. It should never be the one deciding what you charge.
Step five: edit like it is your name on it, because it is. Read every line out loud. Cut anything generic. Fix any bracketed gaps. Correct any claim that drifted from what you can actually deliver. This is the step that separates a proposal that wins from a pile of AI slop, and it is the reason the whole thing still takes thirty minutes and not three. Editing is now the core skill, and we made that case in depth in our piece on giving AI better instructions.
Step six: drop it into your sending tool and send it today. Paste the finished text into Bonsai or your document tool, attach the e-signature, and send it before the day ends. The whole advantage of this workflow is speed, so do not let the proposal sit in a draft folder overnight. A sent proposal that is ninety percent perfect beats a perfect one that never leaves your laptop.
Step seven: queue one follow-up. Set a reminder, or use your tool’s built-in nudge, to follow up in three days if you have not heard back. Most proposals are won on the follow-up, not the send.
The mistakes that will bite you
A few gotchas are worth naming so you can avoid them. Never let the AI invent scope or numbers. If you do not give it a price, a good assistant will happily make one up, and a made-up price in a real proposal is a disaster. That is why the prompt in step three explicitly forbids it. Never send without reading every word, because AI occasionally states a deliverable with total confidence that you never agreed to. And resist the urge to make it longer. Clients skim. A tight two-page proposal closes better than an impressive six-page one nobody finishes.
What success looks like
You will know this is working when proposals stop being a task you dread and start going out the same day, every time. The signal to watch is your response time, the gap between the call and the sent proposal. Get that under an hour and you will feel your close rate move, because you are now the provider who is easy to say yes to. One solo consultant we profiled used exactly this kind of front-office automation to take on twice the clients without working later, and faster proposals were a big part of it. For a fuller look at how the tools in this space compare, this 2026 rundown of AI tools for freelancers and this guide to AI proposal tools are both worth a read.
Here is your one task for today, and it is the step that makes every future proposal fast: open a document and write your proposal skeleton, the six or seven sections you want every proposal to have. Do only that. The next time a discovery call ends well, you will have the mold ready, and thirty minutes later the proposal will be in the client’s inbox while they are still excited they talked to you.
Related reading
- How to Build a Client Onboarding Flow With AI in One Afternoon
- How a Solo Consultant Took On Twice the Clients Without Working Later
- ChatGPT vs Claude vs Gemini: Which $20 AI Should a Solo Owner Actually Pay For?
- You Do Not Need Fancy AI, You Need Better Instructions
What is the longest a proposal has ever sat in your drafts before you sent it? Share your record below, and whether the deal survived the wait.
