The fundraiser's AI prompt library: smarter prompts for charities and nonprofits

Most fundraisers now use AI in some capacity, but few are realizing its full potential.

The problem usually isn't the tool — it's the prompt. Ask ChatGPT to "write me a fundraising email", and you'll get exactly what you asked for: a generic, slightly hollow fundraising email that sounds like every other one. The magic isn't in the asking. It's in how you ask.

A good prompt does the same thing as a good brief. It gives the AI a role, the context it needs, a clear constraint, and a defined output. Do that, and it stops being a novelty and becomes a genuinely useful thinking partner — for audience insights, donor journeys, testing, reporting, and much more.

Below is a library of the prompts I actually find useful, organised by the work that matters. Steal them, adapt them, make them yours. First, read the two rules — they’re the difference between AI that helps and AI that quietly creates a problem.

Two rules before you paste anything in

Rule one: never put sensitive or personal data into a public AI tool. No real supporter records, no donor names or contact details, no anything that could identify a beneficiary, no confidential financial information. Public tools like the free tiers of ChatGPT may use what you type to train future models, and feeding them personal data risks breaching your data protection obligations and your supporters' trust. Anonymise and aggregate before you paste. If you need to work with real data, that's a job for a properly governed, enterprise-grade environment — not a browser tab. For an organisation with "moral" anywhere near its values, this isn't optional housekeeping; it's the whole point.

Rule two: structure beats cleverness. You don't need magic words. The four things that consistently lift the quality of a response are: a role (who the AI should act as), context (the situation and what you know), a constraint (length, tone, audience, brand rules) and an output (the exact format you want back). Every prompt below is built that way, with [bracketed placeholders] for you to fill in. The takeaway is simple: that structure is the actual skill — once you've got it, you can write your own for anything.

One more thing that applies throughout: AI is a first draft and a sounding board, never the final word. Everything it produces — especially anything representing a beneficiary, a community or your brand — needs a human to check it for tone, dignity and accuracy. With bias baked into these models, that review isn't a nice-to-have.

Prompts for audience insight and segmentation

This is where AI earns its keep first, because a better understanding of who you're talking to improves everything downstream. Use it to interrogate what you already know, not to invent supporters out of thin air.

You are an audience insight strategist working with a [cause] charity. Below is a description of our supporter base and what we know about them. Identify 3–4 distinct motivational segments. For each one: name it, describe what it cares about, what motivates it to give, what might cause it to lapse, and one testable message angle. Then tell me what data we would need to validate these segments. Here's what we know: [paste]

And one for pressure-testing a persona you already use — because a persona that nobody has challenged is usually a comfort blanket, not an insight:

Act as a critical audience strategist. Here is one of our donor personas: [paste]. Pressure-test it. Which assumptions are unevidenced? Where might it push us toward generic messaging? Rewrite it as a sharper, behaviour-based segment focused on what this person actually does and why, not just their age and postcode.

If you want to go deeper than a prompt can, turning raw audience data into something you can act on is exactly what our audience intelligence work is built for.

Prompts for donor journeys and retention

Acquisition gets the glory; retention pays the bills. AI is genuinely good at spotting friction in a journey you've stared at for so long you can't see it anymore.

You are a digital fundraising strategist. Below is our current [welcome / first-gift / regular-giving] journey, step by step: [paste]. Identify the three weakest points for a first-time donor, explain why each one risks drop-off, and suggest one specific, testable improvement at each, prioritised by likely impact on second-gift conversion.

For win-back, ask it to think in strategies, not subject lines:

Help me design a lapsed-donor reactivation approach for supporters who gave once 12–24 months ago and haven't returned. Give me three distinct strategic angles, not just copy. For each: what it assumes about why they lapsed, the channel mix, and how I'd measure whether it worked.

Prompts for social listening and narrative analysis

When you've got a pile of social posts, comments or open-text survey responses, AI can find the patterns far faster than you can — provided you point it at the narrative, not just the sentiment.

You are a social listening analyst. Below are posts and comments about [issue / campaign]: [paste]. Summarise the dominant narratives, the emotional drivers behind them, and any framing that contradicts how our sector usually talks about this. Flag any reputational or framing risks, and suggest two message angles that would resonate without feeding a harmful narrative.

Prompts for appeals and campaign copy (built to test, not just to send)

Here's the mindset shift: don't ask AI for a version, ask it for variants you can test. The key takeaway is that this single change turns a copy generator into a test-and-learn engine — and testing is how you actually learn what your audience responds to, rather than what your loudest colleague prefers.

You are a direct-response fundraising copywriter. Write four distinct versions of a [subject line / appeal opening / ask], each built on a different psychological driver — for example, urgency, social proof, identity and impact. Label each driver so I can test them cleanly. Audience: [segment]. Tone: [tone]. Constraints: [word limit / brand rules / things to avoid].

A word of caution on this one above all others: copy that represents the people you exist to serve must be reviewed by a human, every time. AI doesn't understand dignity. You do.

Prompts for SEO, content and AI search

As I wrote in What AI Search Means for Purpose-Driven Organisations, the goal of content is shifting from ranking for clicks to being the source an AI trusts enough to cite. You can use AI to help you write content that AI wants to quote:

You are an SEO and AI-search content strategist. Our audience searches for [topic]. Draft an article outline designed to be cited in AI Overviews and AI search results: use question-based headings, give a one-sentence direct answer under each heading, and identify the single most quotable statistic or definition to lead with. Then suggest a title and meta description.

Prompts for cases for support and institutional fundraising

A trust, foundation or institutional funder reads your bid with a cold, evaluative eye. So get a cold, evaluative read before they do:

Act as an institutional funding assessor reviewing a grant application. Here is our draft case for support: [paste]. Evaluate it the way a major trust or statutory funder would: clarity of need, strength of evidence, theory of change, value for money, and risk. Tell me where it's weakest and what a funder would be unconvinced by. Be direct, not polite.

This kind of strategic, funder's-eye scrutiny is exactly the thinking we bring to digital and fundraising strategy — and the prompt is no substitute for it, but it's a sharp place to start.

Prompts for insight-to-action and board reporting

A lot of brilliant analysis dies in a spreadsheet because nobody turned it into a story a decision-maker can act on. AI is a quietly excellent translator from "data" to "decision":

You are a strategy consultant. Turn the performance data below into a three-paragraph narrative a trustee board would understand: what happened, why it matters, and the single decision it should inform. Avoid jargon. Flag the one number that deserves the board's attention. Data: [paste]

The point isn't the prompts. It's the thinking. The takeaway is that prompts only work when they sharpen judgement, not replace it.

Here's the honest truth underneath all of this. None of these prompts will write your strategy for you. The main takeaway is that they'll help you think faster, challenge your assumptions, draft from a stronger starting point and turn analysis into action more quickly. What they won't do is decide what matters, understand your supporters the way you do, or carry the judgement that fundraising in a purpose-driven organisation demands.

Used well, AI gives you back the most valuable thing you have: time to do the thinking only you can do. Used lazily, it produces confident, polished, generic content that quietly erodes the thing that made your fundraising work in the first place.

The fundraisers who'll win with these tools aren't the ones with the cleverest prompts. They're the ones who already know what good looks like — and use AI to get there faster. The takeaway: judgment still matters most.

If you'd like help turning that into a practical approach for your team — from audience insight to a test-and-learn roadmap — drop us a line. It's the kind of thing we love getting our teeth into.