ChatGPT Prompts for Grant Writing: 30 Sections That Win Funding

The short version: a winning grant application is thirty small writing jobs, not one big one, and ChatGPT is good at exactly the small jobs, need statements, budget narratives, logic models, evaluation plans, as long as you feed it your own facts and never let it invent a statistic. Below are prompts for all thirty sections, grouped by what they do in the application, plus the one habit I watched cost a charity two rejections before we fixed it.

Why grant writing is different from every other kind of prompt

Most guides treat grant writing like blog writing with a bigger budget attached. It is not. A panel member reading your application is often reading forty others that week, unpaid or barely paid, on a train, at 9pm, with a scoring sheet in one hand. They are not looking for beautiful prose. They are looking for the exact word that matches their criteria so they can tick the box and move on. That is the uncomfortable part nobody likes to say out loud: your job is not to persuade, it is to make ticking the box effortless.

This is also why I wrote a separate piece on ChatGPT prompts for grant applications that get read correctly by a panel, which focuses on structure and scoring criteria rather than section-by-section content. This post is the companion to that one. Where that piece is about the skeleton, this is about the thirty pieces of flesh that go on it.

The story: a knock-back that became a funded application

I worked with a domestic abuse charity in the West Midlands on a National Lottery Community Fund application for a caseworker post. Their first attempt, written before we met, was rejected. Second attempt, rewritten by a well-meaning trustee, also rejected, no feedback beyond "did not meet the funding priorities." Third attempt is the one we did together.

The problem wasn't the cause, it was obvious and heartbreaking. The problem was that their need statement talked about domestic abuse in the UK generally, with a Home Office statistic from years earlier, when the funder's criteria explicitly asked for evidence of local need. We rebuilt the need statement using their own referral data, 214 women turned away in the previous financial year because there was no capacity, and that single change, one paragraph, moved the application from "does not meet priorities" to funded at just over £42,000 over two years. The writing time for that redraft, using structured prompts rather than starting from a blank page, was about three hours against the eleven hours the second draft had taken. The lesson wasn't better sentences. It was better facts in the right place.

How to use these thirty prompts without sounding like a robot

Every prompt below has bracketed placeholders. Fill them with your real numbers before you send anything to ChatGPT, never after. If you ask the model to "suggest a statistic," it will, and it will sound entirely plausible, and it may be wrong. I have seen a draft claim "1 in 4 people in the UK experience loneliness weekly" with total confidence and no source, when the real, citable figure was different. Always paste your own evidence in and ask ChatGPT to write around it, never to supply it. If you want the wider version of this rule across other jobs, my ChatGPT prompts list covering 60 prompts across 12 professions has the same warning built into every profession, not just charities.

Run each prompt, then read the output aloud. If it sounds like a brochure, cut every adjective that isn't a number. Funders trust numbers. They are trained to distrust adjectives.

Section group A: framing the need

• 1. Problem statement - "Write a 150-word problem statement for a grant to [funder name] about [issue]. Use this local data: [paste stats]. Do not add any statistic I have not given you."
• 2. Statistics and evidence citation - "Turn these raw figures into three sentences suitable for a grant application: [paste data]. Attribute each figure to its source in brackets."
• 3. Target beneficiary profile - "Describe the people this project will serve, using only these characteristics: [age range, location, circumstance]. Keep it factual, no invented quotes."
• 4. Community consultation summary - "Summarise these consultation notes into a 100-word paragraph showing what the community told us they needed: [paste notes]."
• 5. Root cause analysis - "Using this information about our service users [paste], write a short paragraph explaining the underlying cause of the problem, not just the symptom."

Section group B: the organisation case

• 6. Organisation background - "Write a 120-word organisation background for [name], founded [year], serving [number] people annually in [area], using this history: [paste facts]."
• 7. Track record and past outcomes - "Turn these past project results into a track record paragraph for a funding application: [paste numbers and outcomes]."
• 8. Capacity statement - "Write a paragraph showing we have the staff, premises, and systems to deliver this project, based on: [paste team size, qualifications, premises details]."
• 9. Staffing plan - "Draft a staffing plan section listing roles, hours per week, and responsibilities for this project team: [paste roles and hours]."
• 10. Partnerships and collaborators - "Write a partnerships paragraph describing how [partner organisation] will contribute, based on this agreement: [paste details of what they provide]."

Section group C: the plan

• 11. Goals and objectives - "Convert this overall aim into three SMART objectives with numbers and timeframes: [paste aim]."
• 12. Theory of change - "Write a short theory of change paragraph linking these activities to these outcomes: [paste activities] leading to [paste outcomes]."
• 13. Activities and methodology - "Describe the step-by-step delivery method for this project in 200 words, based on these activities: [paste list]."
• 14. Timeline - "Turn this list of milestones into a 12-month timeline table with dates: [paste milestones]."
• 15. Innovation and differentiation - "Write two sentences explaining what makes this approach different from existing services, based on: [paste what's already out there and how this differs]."

Section group D: money and proof

• 16. Budget narrative - "Write a budget narrative explaining these line items in plain English for a non-financial reader: [paste budget lines and amounts]."
• 17. Budget justification for individual lines - "Justify this specific budget line, [item and cost], in two sentences that show it is necessary and reasonably priced."
• 18. Matched funding statement - "Write a paragraph listing our matched or in-kind contributions to this project: [paste sources and values]."
• 19. Sustainability plan beyond the grant - "Write a sustainability section explaining how this project will continue after the grant period ends, using this plan: [paste future funding or income plan]."
• 20. Risk assessment - "Turn this list of project risks and mitigations into a table format suitable for a grant application: [paste risks and mitigations]."

Section group E: measuring it

• 21. Outcomes and outputs - "Separate these results into outputs (what we did) and outcomes (what changed for people), based on: [paste raw results]."
• 22. Logic model narrative - "Write a short narrative version of this logic model, describing inputs, activities, outputs and outcomes in one paragraph: [paste logic model]."
• 23. Evaluation plan - "Describe how we will measure success for this project, using these tools: [paste surveys, case studies, data tracking methods]."
• 24. Data collection methods - "Write a paragraph describing exactly how and when we will collect this data: [paste what and how often]."
• 25. Dissemination plan - "Write a short section on how we will share learning from this project, based on: [paste channels, e.g. annual report, local press, partner network]."

Section group F: wrapping the application

• 26. Executive summary - "Write a 100-word executive summary of this grant application covering who, what, how much, and why it matters, using this full draft: [paste finished application]."
• 27. Cover letter - "Write a short cover letter to [funder name] introducing this application, referencing their stated priority of [priority] and how our project meets it."
• 28. Letter of intent - "Draft a one-page letter of intent for [funder] summarising the project, the amount requested, and the timeline, based on: [paste key facts]."
• 29. Equity and inclusion statement - "Write a short paragraph describing how this project will reach underrepresented groups, using this outreach plan: [paste specific outreach steps, not generic statements]."
• 30. Closing paragraph and call to action - "Write a two-sentence closing paragraph that restates the amount requested and the outcome it will fund, no fluff."

The bit most grant guides leave out

Here is the uncomfortable truth. Most applications don't get rejected because of weak writing. They get rejected because the project doesn't fit the fund's actual priorities, and no amount of prompting fixes an eligibility mismatch. I have watched organisations spend eight hours polishing a need statement for a fund that had already told them, in the guidance notes, that they only fund capital projects, not staff salaries. Read the guidance twice before you open ChatGPT once. The prompts in this post make your case sharper. They cannot make an unsuitable case suitable.

The second thing worth saying plainly: many funders now ask, directly or indirectly, whether AI was used to write the application. Be straight about it. Using ChatGPT to structure and tighten your own facts is not the same as submitting invented content, and the distinction matters to a reader as much as it matters to a fund's terms and conditions. I wrote about where that line sits for academic writing in how much AI writing is acceptable in a research paper, and the same principle transfers almost exactly to grant applications, you can use AI for structure, clarity, and pace, but every fact, figure, and quote has to be yours.

What changes when you use these prompts

With the domestic abuse charity, the change wasn't stylistic. It was that thirty separate small writing decisions, each backed by a real fact, added up to an application a panel could score quickly. I've since used the same section-by-section approach with a small arts charity applying to Arts Council England and a community food project applying to a local authority discretionary fund, both funded on the first or second attempt after previously stalling. If you want the fuller story of what worked and what didn't across those projects, including which AI tools handled budget tables better than ChatGPT, I wrote it up in what AI tools are useful for writing grant applications, and what I learned helping a charity use them.

One more practical number worth keeping in your back pocket: a full application built section by section using prompts like these, with your own data ready in advance, typically takes a competent writer two to four hours for a mid-size grant, £5,000 to £50,000, versus the eight to fifteen hours I regularly saw before we structured the process this way. The saving isn't the writing itself. It's not having to stare at a blank need statement wondering where to start.

I keep every related walkthrough in the ChatGPT Prompts and Limits: 30 Guides to Fun Prompts, Photo Edits and Daily Caps. To write your own, try the free prompt generator.

Frequently asked questions

Can I copy and paste ChatGPT's grant writing output straight into an application?

No, not without editing. Always check every statistic against your source, remove any generic phrase that could apply to any charity, and read the finished section against the funder's exact wording to make sure you have used their terminology, not ChatGPT's.

Will funders reject an application if they know AI was used to write it?

Most won't, as long as the facts are true and yours. Some funders now ask directly, in which case answer honestly that AI helped structure and edit, while the evidence and figures came from your own records.

How long should each of these thirty sections be?

Most sections should run 80 to 200 words. Executive summaries and closing paragraphs should be shorter, 50 to 100 words. Budget narratives and evaluation plans can run longer if the funder's form gives you the space, but panels reward clarity over length every time.

What is the single biggest mistake people make with ChatGPT and grant writing?

Letting the model invent statistics or examples because a real one wasn't handy. It will produce something confident and specific-sounding that is not true. Always supply your own evidence and instruct the model not to add anything you haven't given it.

Where to check the details

• OpenAI's own platform documentation