The difference between a useful AI prompt and a useless one is specificity. “Write me a freelance proposal” is a useless prompt, it produces a generic draft that takes more time to fix than a blank page would have. The prompts below are specific enough to produce output you can actually use.
These aren’t prompts to generate a full proposal from scratch. They’re task-level prompts that improve specific sections, the three sections that determine whether you close the deal, and five more that add up to a noticeably stronger document overall.
The 8 prompts
1. Rewrite the opening to lead with the client’s situation, not your credentials
Situation: Your draft starts with “I’m a [X] with [Y] years of experience.” That sentence is about you. The opening should be about them.
The prompt:
Rewrite this proposal opening so it leads with the client's specific situation instead of my background. The client's situation: [paste what they told you in the discovery call]. Keep it under 3 sentences. My current opening: [paste your current opening]What good output looks like: A first sentence that could only have been written for this client, something that references the specific problem or goal they described.
What to edit before using: Make sure the output still sounds like you. If you write casually, remove any formal language AI adds. “It is my understanding that your organization seeks to…” is not how humans talk.
2. Generate 3 pricing tier names for a specific project
Situation: You want to offer tiered options but can’t name them without sounding cheesy. “Basic,” “Standard,” and “Premium” signal commodity, not expertise.
The prompt:
Create 3 pricing tier names for a [project type] proposal. The client is a [describe client]. Avoid generic names like Basic/Standard/Premium. The three tiers are approximately $[X], $[Y], $[Z]. Each tier name should signal what's different about it, not just that it costs more.What good output looks like: Names that reflect the actual scope difference between tiers, something like “Foundation,” “Growth,” and “Full Buildout” for a branding project, not “Starter,” “Pro,” and “Enterprise.”
What to edit before using: Pick the names that fit the project. Don’t use all three suggestions if they don’t all work.
3. Expand a thin case study into a persuasive paragraph
Situation: You have a client win you’ve been describing in one sentence. “I helped a brand increase traffic.” That’s not a case study, it’s a sentence.
The prompt:
Expand this case study into 2 persuasive paragraphs. Paragraph 1: The problem and why it was hard. Paragraph 2: My approach and the outcome. Keep it specific. My one-line version: [describe client situation + what you did + measurable outcome]What good output looks like: A paragraph that describes the problem in terms the prospective client recognizes, and a second paragraph that explains the approach in enough detail to demonstrate expertise.
What to edit before using: Replace any vague result language (“significant improvement,” “strong growth”) with the actual numbers if you have them. Specificity is what makes case studies persuasive.
4. Write a risk reversal for a skeptical client
Situation: You’re quoting a project where the client has expressed concern about not seeing results, or where the previous vendor left them burned.
The prompt:
Write a 2-sentence risk reversal for the following proposal situation: [describe the project, the client's stated concern, what you're offering]. The risk reversal should reduce the client's perceived downside without giving away the project for free.What good output looks like: A specific statement about what happens if the work doesn’t meet the agreed criteria, not “I’m committed to your satisfaction,” which means nothing, but something like “If the first revision round doesn’t address all the feedback points we agree on up front, I’ll do an additional round at no charge.”
What to edit before using: Make sure you can actually deliver on whatever the AI writes. Vague AI risk reversals are worse than no risk reversal.
5. Review the scope section for vague language
Situation: You want a second opinion on whether your scope is tight enough. Scope disputes start in the proposal, not during the project.
The prompt:
Review this scope of work section. Flag any deliverable that's described vaguely enough for the client to imagine it includes more than I intend. Also flag any deliverable that's so specific it sounds defensive. [paste scope section]What good output looks like: A marked-up version of your scope with specific flags, “this could be read as including unlimited revisions” or “this revision limit feels like it’s anticipating a problem.”
What to edit before using: This prompt produces a critique, not a rewrite. You’ll still need to revise the flagged sections yourself.
6. Rewrite the project timeline to feel realistic, not rushed
Situation: Your timeline looks tight on paper, or it’s so padded it looks like you’re padding. Neither builds confidence.
The prompt:
Rewrite this project timeline to feel realistic and build confidence. Add a 2-sentence explanation of why each milestone takes the time it does. Current timeline: [paste your timeline]. Client context: [describe the type of client and their urgency level]What good output looks like: A timeline where each phase has a brief rationale, “Discovery and strategy (Week 1): This includes reviewing your existing materials and a 60-minute kickoff call to align on objectives before any design begins.” Clients who understand why each phase takes what it takes are less likely to pressure you to compress it.
What to edit before using: Verify the AI’s rationale actually matches your process. If it invents a phase you don’t do, remove it.
7. Write the “why us” section without sounding like every other proposal
Situation: Your credentials section says “we’ve helped 50+ clients” and nothing else. That sentence exists in half the proposals your client is reading.
The prompt:
Write a "why us" section for a proposal to [client type]. It should be 2 paragraphs. Paragraph 1: What makes our approach specifically right for this type of project (not generic credentials). Paragraph 2: One specific client situation that's directly analogous to what this client is trying to do. My relevant background: [paste your actual experience]. Don't use the phrases "passionate about," "results-driven," or "proven track record."What good output looks like: A first paragraph that describes your approach in terms of what it produces, not what you’ve done. A second paragraph with a named (or anonymized) client situation that mirrors the current client’s challenge.
What to edit before using: Remove any credentialism AI adds back in (“with extensive experience in”). Replace with specifics.
8. Generate the FAQ section from the client’s likely questions
Situation: You don’t know what objections the client will raise, and you don’t have a standing FAQ section.
The prompt:
Write a 4-question FAQ section for a proposal for [describe client and project type]. Questions should be the 4 things a skeptical [client type] would ask before approving a [project type] project. Answer each question directly in 2–3 sentences. Don't use hedging language, give real answers.What good output looks like: Four questions that a real client would actually ask, not “What is your process?” (too generic) but “What happens if the first draft isn’t what we envisioned?” or “How do you handle feedback from multiple stakeholders?”
What to edit before using: Remove any hedging the AI adds. “We strive to…” and “We aim to…” are hedges. Replace with declarative statements.
A note on using the outputs
Treat every AI output as a first draft, not a final draft. The best proposals are roughly 40% AI, structure, first-draft prose, FAQ, and 60% human: the opening, the case study, the pricing rationale.
That ratio shifts toward human as the deal gets bigger. A $2K project proposal can lean heavily on AI-generated structure with light edits. A $50K proposal should feel like it was written specifically for the person reading it, because at that level, the client is evaluating judgment, not just capability.
The prompts above work because they’re editing prompts, not generation prompts. They take something specific you’ve written and make it better, which is a much narrower task than asking AI to write a proposal from a blank page. Narrower tasks produce better output.
Ready to send stronger proposals?
Build, send, and track proposals in one place so follow-up is easier.
Start your free trial →
