Build a proposal kit once (approved company description, service descriptions, proof points, pricing rules and terms), then for each bid give the AI your discovery notes and the client's brief and let it draft the tailored sections: their problem, your solution, the plan. You set the price and check every claim. A day's work becomes about two hours.
The time saving doesn't come from the AI writing more words. It comes from never starting from a blank page and never retyping the same boilerplate. Most of any proposal is the same from bid to bid; the part that wins the work is the page showing you understood this client's problem better than the other bidders did. That page is where the saved hours should go, and it's the page AI can't write without good notes from you.
Where the hours in a proposal actually go
Before changing anything, time your next proposal. Here's the breakdown for an illustrative courier firm bidding for scheduled delivery contracts, before and after moving to a reusable kit and AI drafting:
| Task | Before | With AI and a kit |
|---|---|---|
| Rereading the brief and call notes | 45 min | 15 min (AI summary, then skim) |
| Finding and adapting old proposal text | 1 h 30 min | 0 (kit blocks) |
| Writing the tailored sections | 2 h 30 min | 40 min (draft plus edit) |
| Pricing | 1 h | 30 min (spreadsheet) |
| Formatting and layout | 1 h | 10 min (template) |
| Review and corrections | 45 min | 30 min (checklist) |
| Total | 7 h 30 min | 2 h 5 min |
Notice where the time went before: nearly two and a half hours on hunting for old text and formatting, work that adds nothing for the client. And notice that review time barely falls. AI drafts need checking as carefully as a junior colleague's.
Build the proposal kit once
The kit is a set of approved, reusable blocks that the AI draws on instead of inventing. Put it in one place the AI can read: a ChatGPT Project, a Claude Project or a Gemini Gem (becoming a skill from November 2026) with the files attached and standing instructions. (OpenAI is retiring custom GPTs, so don't build a new one; Projects do this job.) The kit takes four to six hours to assemble the first time, mostly from your best past proposals.
- Company description in three lengths: one line, one paragraph, half a page.
- Service blocks: one per service, describing what's included, what isn't, and how it runs day to day.
- Proof points: real, checkable facts only (years trading, fleet size, clients you have permission to name, measured performance figures with their source and date).
- Method and onboarding: how a new contract starts, week by week.
- Risk and contingency: what happens when a van breaks down or a driver is off sick.
- Standard terms: payment terms, notice periods, liability wording, as approved by whoever looks after your contracts.
- Two or three of your best past proposals, with client names removed, as examples of tone and structure.
- A style card: how you write, words you avoid, spelling conventions.
A filled-in service block from the courier firm's kit:
SERVICE BLOCK: Scheduled multi-drop route
What it is: a fixed daily route, same driver where possible,
collecting from one site and delivering to up to 45 drops.
Included: named route driver and a trained backup; proof of
delivery by photo and signature in the app; exceptions
reported to the client by 11am; weekly delivery report.
Not included: temperature-controlled transport; goods over
25kg per item; weekend runs (available by quote).
How it runs: vans loaded at the client's dock from 6.30am;
route order agreed at onboarding and reviewed monthly.
Proof: [measured on-time rate - insert current figure and period]The bracketed proof line is deliberate. Proof points must come from your records, dated, and the kit should force you to fill them in rather than let the AI make one up. The wider job, setting up the project so every draft starts from your facts rather than the AI's assumptions, is worth doing properly before the first bid.
Capture discovery in a form the AI can use
The quality of a tailored proposal depends almost entirely on the notes from your conversations with the client. Use the same template every time, and fill it in during or right after the call. Here's the courier firm's, filled in for an illustrative wholesaler of pharmacy and healthcare supplies:
DISCOVERY NOTES
Client: healthcare supplies wholesaler, 1 warehouse,
delivers to ~40 independent pharmacies daily
Their problem (their words): "Our own van can't do it all
before the pharmacies open, and when it breaks down we have
pharmacists ringing us all morning."
Current process: 1 in-house van + ad hoc courier bookings
on busy days; ~6 ad hoc bookings a week
What failure costs them: pharmacies switch supplier after
repeated late deliveries (lost 2 accounts this year)
Must-haves: all drops before 9am; proof of delivery;
a named contact; cover when a van is off the road
Nice-to-haves: a second afternoon run twice a week
Decision makers: operations director (decides), finance
director (signs off budget)
Criteria they named: reliability first, then price
Budget signals: currently spends roughly $6,000 a month on
the van, driver and ad hoc bookings combined
Timeline: decision by end of next month; start in 6 weeks
Risks they raised: handover period; drivers not knowing
the pharmacies' back-door access"Their words" is the most useful line on the page. When a proposal repeats the client's own description of the problem, they know you listened.
Drafting section by section
Asking for a whole proposal in one prompt produces a long, generic document that takes longer to fix than to write. Draft one section at a time, in this order: understanding of the problem, solution, plan and onboarding, risk, and the executive summary last, when everything it summarises exists.
The understanding section
Using DISCOVERY NOTES and the client's brief, write the
section "Your situation and what you need" (250-350 words).
- Open with their problem in their own words.
- Show the cost of the problem using only facts in the notes.
- List their must-haves exactly as stated.
- Don't mention our company or services yet.
- Don't add facts, figures or assumptions not in the notes.
Follow the style card.An illustrative first draft opened: "In today's fast-moving healthcare sector, timely delivery is more critical than ever. Your pharmacies rely on you, and every late delivery risks patient care and your reputation…"
That's three sentences of nothing, and "risks patient care" is a claim the client never made. After cutting it and adding "Start with the quote from Their problem" to the prompt, the second draft opened: "Your van can't reach all 40 pharmacies before they open, and when it breaks down, as you put it, 'we have pharmacists ringing us all morning'. Around six ad hoc courier bookings a week fill the gaps, and late deliveries have already cost you two pharmacy accounts this year." That's the paragraph the operations director will read twice.
The solution and the plan
Give the AI the relevant service blocks and the notes, and ask it to match each must-have to a specific part of the service, in a table if the client's brief is formal. The courier firm's draft mapped "all drops before 9am" to a 6.30am loading slot and route plan, "cover when a van is off the road" to its backup-van arrangement, and "drivers not knowing back-door access" to a two-week onboarding in which the current in-house driver rides along. That last point came straight from the risks line in the notes, and it's the kind of detail that makes a proposal feel written for one client.
The executive summary, last
Only when every other section is final, ask for a half-page summary of the problem, the recommended option, the price and the start date, using only what's in the finished sections. Check the price and dates in it against the pricing table; summaries are where mismatched numbers hide.
Pricing: keep the numbers out of the AI's hands
Language models make arithmetic slips and can't be trusted to apply your rate card consistently. Price in a spreadsheet, then let the AI explain the options in words. The courier firm's three options, calculated from its rate card (illustrative rates):
| Option | What's included | Calculation | Monthly price |
|---|---|---|---|
| A: Morning route | One van, all drops before 9am, weekdays | 1 van × 7 h × $44/h × 22 days | $6,776 |
| B: Morning route plus two afternoon runs | As A, plus afternoon runs Tue and Thu | $6,776 + (3 h × $44 × 9 days) | $7,964 |
| C: Morning route plus on-demand urgents | As A, plus same-day urgent drops | $6,776 + $26 per urgent drop | $6,776 plus usage |
The prompt for the pricing section then says: "Explain options A, B and C using the table exactly as given. Don't change or round any figure. Say which we recommend and why, using the client's criteria." Against the client's current spend of roughly $6,000 a month, option A costs more but removes the breakdowns and ad hoc bookings. A good draft says that honestly rather than pretending it's cheaper. For the reasons AI struggles with sums, and what to check, see catching made-up figures in AI-drafted proposals.
The courier firm's bid, start to finish
Here's how the two hours were spent on the healthcare wholesaler's proposal:
- 15 minutes: paste the brief and discovery notes into the proposal project; ask for a one-paragraph summary and a list of anything the brief asks for that the notes don't cover. It flagged that the brief asked for insurance details, which weren't in the notes or kit; the owner added the firm's actual cover from its policy schedule.
- 40 minutes: draft and edit the understanding, solution, plan and risk sections.
- 30 minutes: price the three options in the spreadsheet and draft the pricing explanation.
- 10 minutes: executive summary and assembly into the branded template.
- 30 minutes: the review pass below, with a colleague reading it cold.
The firm now sends proposals within two working days of the discovery call instead of a week, and can bid for smaller contracts that weren't worth a full day's writing before. Whether faster proposals win more often is something to measure over a quarter, not assume.
The review pass: errors AI adds to proposals
Read every AI-assisted proposal against this list before it goes out:
- Another client's name or details. Reused examples carry old names. Search the document for every past client name in your kit.
- Figures you can't back up. "A 99.8% on-time rate" when you've never measured it. Every number needs a source you could show the client.
- Promises beyond your service. "Temperature-controlled where required", when your vans aren't. Compare every capability claim with the service blocks' "Not included" lines.
- Numbers that don't match. The summary says $6,776; the table says $6,766. Check every price and date against the pricing sheet.
- Changed legal wording. AI "tidies" standard terms. Paste terms in unchanged from the kit, never regenerated.
- Filler and hype. "In today's fast-moving sector", "state-of-the-art", "we pride ourselves". Cut on sight.
- Things the brief asked for that are missing. Tick off every requirement in the brief against a heading in the proposal.
A realistic slip from the courier firm's early use: a proposal to a garden centre chain went out with a paragraph about "your 40 pharmacy customers", carried over from the healthcare bid through the project's chat history. The fix was simple: start each proposal in a fresh conversation inside the project, so the kit is shared but the previous client's notes aren't. The AI proposal mistakes that cost you the job lists more of these, and formal tenders with scoring criteria have their own extra rules, covered in responding to tenders and RFPs with AI.
A shorter version for small contracts
Not every bid deserves ten pages. For small or low-risk work, the same kit produces a one-page letter proposal in about thirty minutes. The courier firm's example: an independent florist asked for two scheduled deliveries a week to a handful of wedding venues and offices. The owner pasted three lines of notes into the project and asked for a one-page proposal using the company description (one-line version), the scheduled-route service block, the standard terms and a single price from the spreadsheet.
Write a one-page proposal letter (under 300 words) for NOTES.
Structure: their need in one sentence; what we'll do (from the
scheduled-route block, only what applies); price from PRICE
exactly as given; start date; terms by reference ("our standard
terms, attached"). Plain, warm, no headings.
NOTES: florist, 2 deliveries a week (Tue, Fri), 4-8 drops each,
fragile arrangements, must arrive before 10am, wants to start
next month. PRICE: $185 per run.The draft needed one correction: it promised "careful handling of all arrangements" and then, pulling from the service block, "items up to 25kg", which read oddly for flowers. Cutting the weight line and adding the florist's own concern ("arrangements travel upright in crates we supply") took two minutes. The point is that the kit scales down as well as up; the discipline of notes, approved blocks and spreadsheet pricing is the same.
Keep the kit improving after every bid
A kit that never changes slowly goes stale. Spend fifteen minutes after each decision, won or lost:
- Won: add the winning understanding section to the kit's examples (with the client's name removed), and update proof points if the contract gives you a new, nameable reference.
- Lost: ask the client why, if they'll say, and write the answer in a "lessons" file the project reads. "Price was 12% above the winner" and "they wanted a named account manager" both change the next draft.
- Every quarter: refresh measured figures (on-time rate, fleet size) with their dates, and delete any proof point you can no longer back up.
Which tools to write proposals in
- A general AI assistant with projects (ChatGPT, Claude or Gemini) is enough for most small businesses. It drafts; your usual word processor and template handle the layout. Claude versus ChatGPT for business writing compares the two for longer documents.
- AI inside your word processor is handy if your kit lives in Word or Google Docs. Microsoft 365 Copilot Business costs $21 a user a month on annual billing (with a promotional $18 until the end of 2026), and can draft in Word from your own files; drafting proposals with Copilot in Word shows how. Gemini is built into Google Workspace Business Standard and above.
- Dedicated proposal software such as PandaDoc, Proposify or Qwilr adds branded templates, e-signatures, pricing tables clients can choose from, and alerts when a proposal is opened. Qwilr, for example, starts at $35 a user a month billed annually. It's worth it when you send many proposals a month or need e-signatures; proposal software versus a general AI assistant goes through the trade-off.
Before and after: an executive summary
Before (unguided AI): "We are delighted to present this proposal for your consideration. As a leading logistics provider with a passion for excellence, we offer tailored, reliable and cost-effective delivery solutions designed to meet your unique needs and exceed your expectations."
After (from the finished sections): "Your pharmacies need every delivery before 9am, and your single van can't guarantee that. We propose a dedicated morning route (Option A, $6,776 a month) with a named driver, a backup van and photo proof of every delivery, starting six weeks from sign-off after a two-week ride-along with your current driver. If you'd like the twice-weekly afternoon runs you mentioned, Option B adds them for $7,964 a month in total."
Every sentence in the second version comes from the client's notes or the pricing sheet. That's the test for any AI-drafted proposal: could each line have been written for any other client? If it could, cut it or make it specific.
Further reads
- How to Create Quotes and Estimates in Minutes With AI — For shorter, repeatable jobs where a quote beats a proposal.
- How Consultants Use AI to Write Proposals in Under an Hour — The same approach applied to consultancy proposals.
- How to Write a Sales Deck With AI That Wins Meetings — Turning the proposal into a deck for the pitch meeting.
- How to Mention AI Use in Client Contracts and Proposals — Whether and how to mention AI use in proposals and contracts.
- How to Give AI Your Business Context So Answers Fit — Setting up a project so every draft starts from your own facts.
- How to Build a Brand Voice Guide That AI Can Follow — Making every proposal sound like your company.
- AI Engagement Letters for Accountants: Faster Drafts, Clear Scope — Which parts of an engagement letter AI should draft, five prompts with sample outputs, and the partner checks that keep scope clear and terms untouched.
- How Interior Designers Use AI for Mood Boards and Presentations — How interior designers use AI for concept boards, real-product selection boards and client decks, with prompts, sample outputs and the traps to avoid.
- How Small Agencies Use AI to Write New-Business Pitches — Where AI shortens a small agency's pitch: go/no-go scoring, prospect research, idea stress-tests, case-study matching and a sceptical-buyer rehearsal.
- How Freelancers Use AI for Invoices, Proposals and Chasers — One thread from proposal to payment: AI drafts the scope, checks the invoice and writes the reminders, and your invoicing software sends them on time.
- How to Write Case Studies From Customer Interviews Using AI — Interview questions, transcript-to-brief and draft prompts, and an approval routine for case studies where every result comes from the customer, not the model.
- How to Write Sales Call Scripts With AI That Don't Sound Scripted — Swap word-for-word scripts for an AI-built call map: openers, questions and objection cards drawn from your own best calls, then rehearsed aloud with AI.
- How Small Businesses Use AI in Sales: 10 Real Examples — Ten practical sales workflows, from sorting enquiries to checking proposals, with sample outputs and the decisions staff should keep.
- How to Use Gemini in Google Docs to Draft and Edit Faster — Drafting from @-mentioned files, Refine before and after, Help me create with a template, and the number mistakes to catch before you accept.
- AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.
Sources: OpenAI help pages (custom GPT retirement and Projects); Qwilr pricing page; Microsoft 365 Copilot Business pricing and Google Workspace plan pages from the vendors' pages. Checked September 2026.