Six Figures of Proposals in Two Hours With AI
Two hours.
One recent morning I sat down with three proposals hanging over my head. Together they were worth six figures of work. Anyone who writes proposals knows the drag of it. Pulling scope together, finding the right language, formatting, checking the numbers, and doing all of it while the actual billable work waits.
By the end of the morning all three were drafted, formatted, and checked. Two hours, start to finish.
What I did differently has almost nothing to do with which AI tool I used. It has everything to do with how I treated it.
I opened Claude and I talked to it like it was my proposal manager. Not a search box. A colleague. I told it the role it was playing. I gave it the context a real proposal manager would need. Who the client was, what the scope looked like, what we had proposed before, what format the final document had to land in. Then we went back and forth, the way two people do when they are hammering out a draft together.
We turned out six figures worth of proposals in two hours. It put them in the right format, ready to drop into whatever system you run. Word, Google Docs, PandaDoc. And before anything went out, we checked the work and benchmarked it against what is actually being offered out there in the industry.
Think about a sharp new hire starting at your company on Monday. You would never tell them three words and walk away expecting a finished proposal by lunch. You would brief them. Client, budget, history, what won last time, what the deliverable looks like. Then you would review their draft, push back on the weak parts, and they would revise. Nobody calls that inefficiency. That is just how high quality workflows run.
Most people skip all of that with AI. They type a one-line request into what they treat as a fancy search box, get a mediocre generic answer, and conclude AI won’t work for me. AI is not the problem. You were just missing the systematic workflow, the context, and the feedback.
How do you actually talk to AI like a colleague?
Three key ingredients for your approach.
First, give it a role. “Act as my proposal manager” changes the entire conversation. The model now knows what job it is doing, what a good output looks like, and what questions it should be asking you. This works for any role you can name. Estimator, editor, scheduler, patient explainer of contract language.
Second, load it with context. These tools love information and data. Give more than feels natural. The client, the history, the constraints, the format, an example of a past proposal you were proud of. Every piece of context you hold back is a gap the model fills with a guess.
Third, push back. The first answer is not always the best answer. When something is incorrect, say so, the same way you would to a colleague in the office. This section is too generic. Add more detail here. Here is what actually differentiates us. The second and third passes are where the good work shows up. You have to keep working with it, just like you would with another human.
Does the checking go away?
No, and I want to be direct about that, because it is the part people worry about.
The way I work now is simple. Use AI to plan the work. Use AI to do the work – the calculations, the writing, the first draft of the proposal. Then use AI to check the work. I will take an output and run it through a second model, sometimes a third, and have them peer check each other. ChatGPT reviewing Claude and back again.
Then the expert human in the loop check point and approval remains intact. I check it. My peer checks it. The approver signs off. None of that goes away. What changed is the time it takes to get to that checkpoint. Drafting that used to swallow whole days is ready for review in a morning.
That is the shift I want you to make this week. Stop typing search queries and start briefing them like the most capable colleague you have ever had, one who never slow down, never takes a day off, and accepts unlimited feedback. The judgment, the standards, the final call, those stay yours. Those were always the most valuable part of your work anyway – the taste, judgement, creativity, and decisions.
If you try this on your next proposal, I would like to hear how it goes. Reply and tell me, or follow along here. I share what is working as I find it.