Free · From Brazen Fundraising

The Fundraiser's AI Prompt Library

30 prompts for the work you actually do

Written by a human — Emily Bocking, Brazen Fundraising · no hype

You're the one your organization expects to figure AI out. On top of the appeal, the grant due Friday, the board package, the gala. Nobody handed you a manual. Nobody told you what's safe. And every third post in your feed is a tech bro yelling that you're already behind.

You're not behind. You're busy. Those are different problems.

This is the manual nobody handed you. Thirty prompts for the work that actually eats your week — grant drafts, thank-yous, meeting prep, newsletters, prospect research, the admin that steals your Fridays. Each one is copy-paste ready. Each one tells you why it works. And each one tells you, where it matters, what never goes in.

Because here's the thing the demos skip: the tools are not your problem. The problem is that nobody showed you how to use them without putting your donors at risk. That's the whole reason this exists.

Read the ground rules first. They're the difference between AI that saves you four hours a week and AI that lands your organization in a data-privacy conversation you don't want to have. Then use whatever you need, in whatever order your week demands.

Before you paste anything: the ground rules

Three rules. Read them once, and they'll keep you safe for good.

The Dignity Line — what never touches a free tool

There's a line, and it doesn't move. On one side: the work. On the other side: your donors' dignity.

Donor personal information never goes into a free, public AI tool. Not their names. Not their gift histories. Not their contact details. Not their health, their family situations, their divorce, their windfall, the reason they gave, the reason they stopped. None of it.

Free tools like the public version of ChatGPT can use what you type to train their models. That means a donor's name and giving history, pasted into a chat box to "draft a thank-you," has left your control. You can't get it back. And that donor never agreed to it.

So the rule is simple: the AI does the writing, you supply the donor. You draft the thank-you letter with a placeholder — [DONOR NAME], [GIFT AMOUNT] — and you fill in the real details yourself, in your own documents, after. The tool never sees the person. It only sees the shape of the letter.

Every prompt in this library that touches donor work is built this way. Where a prompt could tempt you across the line, there's a Dignity Line note telling you exactly what to keep out. Follow it. This is the one rule that protects the people who trust you.

(One honest caveat: some paid, enterprise AI tools — the ones your organization pays for, with a signed agreement that says your data won't be used for training — are a different conversation. If your org has one of those, ask whoever set it up what's covered. Until you know, treat every tool like a free one. The Dignity Line holds by default.)

Read every word before it goes anywhere external

AI is a fast, confident intern who sometimes makes things up. It will invent a statistic. It will state a "fact" about your organization that isn't true. It will get a donor's tone wrong. It does all of this in flawless, believable prose — which is exactly what makes it dangerous.

So: nothing an AI writes leaves your hands unread. Not a grant paragraph, not a thank-you, not a newsletter line, not a single sentence going to a donor or a funder. You read it word for word. You are the editor. The AI is the draft. The draft is never the final.

This isn't extra work. It's the same read you'd give your own first draft. You're just starting from a fuller page.

The free-tool rule

Assume the tool you're using is free and public unless you know for certain it isn't. Free tools are wonderful for the work in this library — drafting, summarizing, brainstorming, restructuring, tightening. They're genuinely useful, and they'll give you hours back.

They are not a database. They are not a CRM. They are not a safe place to store or process the people you serve. Use them for the shape of the work. Keep the people in your own systems.

That's it. Three rules. Now the prompts.

Grant drafting & reporting

The work that eats whole days. AI won't win you the grant — your program does that. But it'll get you from a blank page to a real draft in a fraction of the time, and it'll turn a finished report into a plain-language version a board can read.

1

Turn program notes into a first grant draft

The task You have messy notes about a program. You need a fundable narrative.

You're a grant writer for a Canadian nonprofit. I'm going to give you rough notes
about a program. Turn them into a clear, compelling first-draft narrative of about
[400] words for a grant application. Use plain, specific language — no jargon, no
inflated claims. Structure it as: the need, what we do about it, who it reaches, and
what changes as a result. Ask me up to three questions if anything critical is missing.

Here are my notes:
[PASTE YOUR PROGRAM NOTES — the problem, the activities, roughly who's served, any
outcomes you already know]

Why this works It gives the tool a role ("grant writer"), a length, a structure, and permission to ask questions instead of inventing answers. That last part is what stops it from making up outcomes you didn't give it.

Dignity Line

Your program notes are fine — those describe your work, not a person. But do not paste the names or personal stories of specific people you serve unless they've given documented consent to have their story used. Anonymize. "A single mother of two" carries the story. Her actual name doesn't need to be in the chat box.

2

Match your program to the funder's priorities

The task A funder's guidelines use their language. Your program uses yours. You need to speak theirs without lying.

I'm applying to a funder. Below are their stated funding priorities, and below that is
a plain description of my program. Show me, honestly, where my program genuinely aligns
with their priorities and where it doesn't. For the real alignments, suggest language
that connects my work to their goals without overstating or inventing anything. Flag
anything that would be a stretch so I don't oversell it.

FUNDER PRIORITIES:
[PASTE THE FUNDER'S PUBLISHED PRIORITIES]

MY PROGRAM:
[PASTE YOUR PLAIN PROGRAM DESCRIPTION]

Why this works Asking it to flag the weak alignments is the trick. Left alone, AI is a people-pleaser — it'll find "alignment" everywhere. Telling it to be honest about the stretch keeps your application credible and keeps you out of a relationship built on a fudge.

3

Draft the answer to a hard application question

The task The question is "Describe your organizational capacity and sustainability." Your brain goes blank.

Help me draft an answer to this grant application question: "[PASTE THE EXACT QUESTION]".
The word limit is [250] words. Here's what's true about us:
[PASTE 4-6 BULLET POINTS OF FACTS — years operating, team size, budget, key partnerships,
track record, funding mix]
Write a confident, specific, honest answer. Don't inflate. If a claim needs a number I
haven't given you, leave a clearly marked [BRACKET] for me to fill in rather than guessing.

Why this works The [BRACKET] instruction is doing quiet, important work — it turns the AI's habit of confidently inventing numbers into a checklist of things you fill in. You never ship a fabricated stat because the tool never wrote one.

4

Translate a finished report into board-plain language

The task You wrote a dense funder report. Now your board wants the short version.

Below is a section of a grant report I wrote for a funder. Rewrite it for a volunteer
board member who doesn't work in the sector and has five minutes. Keep every fact exactly
as written — do not change any numbers or claims. Just make it clear, plain, and short.
Aim for [150] words. Lead with what changed for the people we serve.

[PASTE YOUR REPORT SECTION — with any donor or client names already removed]

Why this works "Keep every fact exactly as written" tells the tool to restructure, not re-report. You get plainer language without risking a number drifting.

Dignity Line

Strip client and donor names out of the report section before you paste it. The board version doesn't need them and the tool shouldn't see them.

5

Build a reusable report outline

The task You write the same kind of report four times a year. You want a skeleton you own.

I write quarterly impact reports for funders. Build me a reusable outline I can fill in
each time — the standard sections a strong, honest impact report needs, with a one-line
prompt under each section reminding me what belongs there. Keep it practical for a
small shop with limited time. No fluff sections we'd only include to look bigger than
we are.

Why this works This one asks the tool for a tool — a template you keep, not a one-off draft. It's the difference between borrowing time once and building yourself an hour back every quarter. And "no fluff sections" keeps it honest to a small shop's reality.

Donor thank-yous & stewardship

The work that matters most and gets rushed most. AI can't feel gratitude. But it can help you get unstuck, vary your language so twelve letters don't read identically, and make sure the donor — not your organization — is the hero of the sentence.

6

Draft a thank-you you'll finish yourself

The task A gift came in. You want to thank them well, not with the same tired template.

Help me draft a warm, specific, non-generic thank-you letter for a donor gift. Do NOT
use the donor's real name or details — I'll add those myself. Use placeholders:
[DONOR NAME], [GIFT AMOUNT], [WHAT IT FUNDS].
The tone: genuine, warm, human — not corporate, not gushing. Make the donor the hero of
the letter, not the organization. About [150] words. The gift funds this kind of work:
[DESCRIBE THE WORK IN GENERAL TERMS — no client names]

Why this works The placeholders are the Dignity Line built into the prompt. The tool drafts the shape; you supply the human. "Make the donor the hero" is the instruction that flips a letter from look what we did to look what you made possible — which is the whole point of a thank-you.

Dignity Line

This is the prompt people are most tempted to break. Do not paste the donor's name to "make it more personal." Fill it in yourself, after. The warmth comes from you, not from the tool knowing who they are.

7

Get unstuck on the "why it mattered" line

The task You know what the gift bought. You're stuck on why it mattered.

I'm writing a thank-you and I'm stuck. A gift funded [DESCRIBE, IN GENERAL TERMS, WHAT IT
FUNDED — no names]. Give me five different one-sentence ways to describe why that mattered
— written for the donor, focused on the change they made possible, not on our activity.
Range from understated to more emotional so I can pick what fits this person.

Why this works Five options, ranging in tone, means you're choosing — not accepting. You know this donor. The tool doesn't. It gives you the raw material; your judgment picks the one that sounds like the relationship.

8

Vary a batch so they don't read like mail-merge

The task Twelve thank-yous for twelve similar gifts. You don't want them identical.

Here's one thank-you paragraph I've written. Give me four alternative versions that say
the same thing with genuinely different wording, rhythm, and opening — so a batch of
letters doesn't read like the same template twelve times. Keep the warmth and keep it
specific. No placeholders needed; there are no names in here.

[PASTE YOUR ONE PARAGRAPH — with no donor names in it]

Why this works Donors talk to each other. Board members compare notes. Four real variations means the person who gave $50 and the person who gave $5,000 don't get word-for-word the same letter. The specificity you keep; the sameness you kill.

9

Draft a stewardship check-in that isn't an ask

The task You want to reach out to a donor just to update them. No ask. That muscle has gone stiff.

Help me draft a short, warm donor update — a check-in, not an ask. The point is to show
a donor the difference their past support is making, with no request for anything.
Use placeholders [DONOR NAME] and [WHAT THEY SUPPORTED]. About [120] words. It should
feel like a note from a person who's glad to keep them in the loop, not a fundraising
touch. End with something human, not a call to action.

Why this works Naming what it's not ("not an ask," "not a fundraising touch") is what keeps the tool from sliding a soft solicitation into the last line out of habit. Stewardship that isn't secretly an ask is how trust gets built. This prompt protects that.

10

Turn a milestone into stewardship prompts

The task Your program hit a milestone. You want to steward donors around it without writing twelve different things.

Our organization just reached this milestone: [DESCRIBE THE MILESTONE — no client names].
Give me three ways I could share this with donors as stewardship (not solicitation):
one for a quick email, one for a social post, one for a personal note to a major supporter.
Keep each donor-centered — frame it as something they helped make happen. Plain and warm,
not triumphant.

Why this works One milestone, three surfaces, all donor-centered. "Not triumphant" is the guardrail against the "we're thrilled to announce" tone that makes stewardship feel like a press release. The donor made it happen; the copy should say so.

Meeting prep & summaries

The invisible time-sink. AI is genuinely good here — turning your own scattered notes into something usable, and turning a long transcript into the three things you actually need to do next.

11

Prep for a donor meeting from your own notes

The task You have a meeting tomorrow and a page of your own scattered notes about the relationship.

I have a donor meeting coming up. Below are my own notes about this relationship and what
I want to get out of the meeting. Help me prepare: give me a short, clear prep sheet —
my three goals for the meeting, two or three good open questions to ask them, and a couple
of things to listen for. Keep it conversational, not scripted. I want to have a real
conversation, not run an agenda.

MY NOTES:
[PASTE YOUR OWN NOTES — see the Dignity Line note below before you do]

Why this works "A real conversation, not an agenda" keeps the output human. It preps you to listen, not to perform. That's the difference between a meeting that builds a relationship and one that reads as a pitch.

Dignity Line

This is the sharpest edge in the whole library, so read carefully. Your prep notes are full of exactly what the Dignity Line protects — the donor's name, their giving history, maybe personal details. Don't paste those. Instead, anonymize before you paste: "a long-time donor, gives annually, recently mentioned an interest in [cause area], seemed hesitant last time about [general topic]." That gives the tool everything it needs to help you prep, and it gives the donor everything they're owed: their name stays yours. If that feels like too much effort for a busy Tuesday — that effort is the job. It's what keeping trust actually costs.

12

Turn a meeting transcript into next steps

The task You recorded a team meeting (with everyone's OK). Now you have a 40-minute transcript and no time.

Below is a transcript of an internal team meeting. Pull out only what's useful:
1. Decisions we made
2. Action items — who owns each one, and any deadline mentioned
3. Anything left open or unresolved that needs a follow-up
Ignore the chit-chat. Be concise. If an owner or deadline wasn't clearly stated, say
"unassigned" rather than guessing.

TRANSCRIPT:
[PASTE THE INTERNAL MEETING TRANSCRIPT]

Why this works "Say unassigned rather than guessing" stops the tool from inventing accountability that nobody agreed to in the room. You get a real action list, not a plausible-sounding fiction.

Dignity Line

This works for internal meetings about operations. If a meeting discussed specific donors or clients by name and personal circumstance, strip those names before pasting, or don't use a free tool for that one at all. A transcript of you and a donor talking about their late husband is not something a free tool should ever see.

13

Summarize a long document you have to act on

The task A funder sent a 20-page guidelines PDF. You need the parts that affect your application.

Below is a funder's guidelines document. I'm applying. Give me:
1. Eligibility — am I clearly in, clearly out, or is it unclear?
2. What they fund and what they explicitly won't fund
3. Deadlines and key dates
4. Anything unusual or easy to miss (word limits, required attachments, reporting terms)
Quote the exact wording for anything about deadlines or eligibility so I can double-check it.

[PASTE THE GUIDELINES TEXT]

Why this works "Quote the exact wording" for the make-or-break details means you're not trusting a paraphrase on the things that would sink your application. You verify against the source. The tool saves you the reading; it doesn't get the final say.

14

Draft an agenda that keeps a meeting short

The task You're running a meeting and you want it to end on time for once.

Help me build a tight agenda for a [45]-minute meeting. The purpose is: [STATE THE ONE
THING THIS MEETING MUST ACHIEVE]. Here's who's coming and what each needs from it:
[LIST ATTENDEES AND WHAT THEY NEED — roles, not personal details]
Give me a time-boxed agenda that protects the purpose, cuts anything that could be an
email, and ends on time. Flag anything on here that doesn't actually need a meeting.

Why this works "Flag anything that doesn't need a meeting" turns the tool into the person at the table brave enough to say "this could've been an email." It defends your calendar, which nobody else is doing for you.

15

Prep talking points for a board conversation

The task You have to raise something hard with your board. You want to be clear, not defensive.

I need to raise a difficult topic with my volunteer board and I want to be clear and calm,
not defensive. The topic: [DESCRIBE THE ISSUE IN PLAIN TERMS]. Here's the context they'll
need: [BRIEF CONTEXT]. Help me prepare talking points that: state the situation honestly,
explain why it matters, and offer a clear ask or recommendation. Keep it respectful and
non-blaming — I have to keep working with these people. Anticipate two likely pushbacks
and help me prepare an honest response to each.

Why this works "I have to keep working with these people" tells the tool to draft for the relationship, not just the argument. Anticipating pushback means you walk in prepared instead of blindsided. This is prep for a hard conversation you'd otherwise lose sleep over.

Newsletters & comms

The recurring deadline that never stops coming. AI won't give your newsletter a soul — your stories do that. But it'll get you past the blank screen, tighten what's baggy, and turn one story into the three formats you need it in.

16

Get past the blank newsletter screen

The task The newsletter is due and you've got nothing but a vague sense of what happened this month.

Help me brainstorm this month's donor newsletter. Here's roughly what happened at our
organization this month: [DUMP EVERYTHING — programs, milestones, events, a good moment,
a challenge — no client or donor names]. Suggest three or four short newsletter items
from this, each with a plain-language headline and a one-line description of the angle.
Lead with what a donor would actually care about, not internal news. Tell me honestly
if something on my list isn't worth a donor's attention.

Why this works "Tell me honestly if something isn't worth their attention" is the editor you don't have. It stops the newsletter from becoming a list of internal updates nobody outside the office cares about. Donors want to know what their money did — not that you hired a new coordinator.

17

Tighten a draft that's too long

The task You wrote a newsletter piece. It's 400 words. It should be 180.

Below is a newsletter piece. Cut it to about [180] words without losing the heart of it or
any specific, concrete detail. Keep the human warmth. Cut throat-clearing, repetition, and
anything a donor would skim past. Show me the tightened version, then tell me in one line
what you cut and why so I can put anything important back.

[PASTE YOUR DRAFT — no donor or client names]

Why this works "Tell me what you cut and why" means you stay in control of the edit. AI is good at cutting words and bad at knowing which sentence carries the soul. Making it show its work lets you rescue the one line it shouldn't have touched.

18

Write three subject lines that aren't clickbait

The task Your newsletter's ready. The subject line is blank and you hate all your ideas.

Write me [6] email subject lines for a donor newsletter. The main story is: [ONE SENTENCE
ON THE LEAD STORY]. I want subject lines that are honest and human — no clickbait, no fake
urgency, no ALL CAPS, no "you won't believe." Just clear and inviting enough that a busy
supporter opens it. Give me a range from plain to warmer.

Why this works Banning the manipulation tactics by name ("no clickbait, no fake urgency") gets you subject lines that respect the donor. Your list opts in because they trust you. A bait-and-switch subject line spends that trust for one open.

19

Turn one story into three formats

The task You have one good story. You need it as a newsletter blurb, a social post, and a line for the annual report.

Here's a story from our work: [DESCRIBE THE STORY — anonymized, no real names, consent
respected]. Give me three versions of it:
1. A short newsletter item (about 100 words)
2. A social media caption (short, warm, no hashtags — I'll add those)
3. One or two sentences for our annual report
Keep the human core in all three. Don't sensationalize or turn the person into a prop.

Why this works One story, three surfaces, one draft session — that's real time saved. "Don't turn the person into a prop" is the guardrail that keeps a client's story from becoming poverty tourism. The people in your stories are people, not content.

Dignity Line

Only use a client or beneficiary story if that person gave real, documented consent to have it shared — and even then, anonymize in the prompt. The tool doesn't need their name to help you write it well.

20

Draft a plain-language explainer of your work

The task Someone always asks "so what does your org actually do?" and you fumble it.

Help me write a clear, jargon-free explanation of what my organization does, for someone
who's never heard of us and has thirty seconds. Here's the plain version of our work:
[DESCRIBE WHAT YOU DO]. Give me three lengths: one sentence, one short paragraph, and a
three-sentence version. No sector jargon, no acronyms, nothing that needs insider
knowledge to understand. A smart person outside our world should get it immediately.

Why this works Three lengths means you're ready whether you've got a sentence or a paragraph — the elevator, the email signature, the "tell me about your work" at a party. Killing the jargon is what lets an outsider actually understand, which is the entire job of an explainer.

Research & prospect prep (without donor PII)

This is the one everyone gets wrong. The temptation is to paste a prospect's name and ask the AI to "tell me about them." Don't. Here's the safe pattern that gets you real value without ever putting a person into a tool.

The safe pattern: public info only, and you do the looking

Read this before the prompts. It's the whole method.

You do not paste a real prospect's name into a free AI tool and ask it to research them. Two reasons. First, the Dignity Line — you'd be feeding a private individual's identity into a public tool. Second, it doesn't even work well: the tool will confidently make things up about a real named person, and now you're prepping off a fiction.

Instead, you use AI for what it's actually good at in research: helping you know what to look for, and helping you make sense of what you find yourself. You do the looking — in public sources, in your own CRM, in a real conversation. The AI helps you prepare and organize. It never becomes the source, and it never holds the person.

21

Build a research checklist for a prospect meeting

The task You have a first meeting with a potential major donor. You want to prep well and ethically.

I have a first meeting with a potential major donor. Help me build an ethical, practical
research checklist — the kinds of PUBLIC information that genuinely help me prepare for a
respectful, relevant conversation, and where I'd typically find each (public sources only).
Then give me a short list of things I should NOT go looking for because they'd cross a line
into someone's private life. I want to be prepared, not invasive.

Why this works It turns the tool into a research coach instead of a research database. You get a checklist you then work through yourself, in public sources — and a clear line about what not to dig into. Prepared, not creepy.

Dignity Line

Notice there's no name in this prompt, and there never should be. You're asking how to research, not asking the tool to do the research on a named person. That distinction is the whole safe pattern.

22

Role-play a donor conversation (with a made-up donor)

The task You want to practice a nervous ask out loud, but you don't want to practice on the real relationship.

I want to practice a major gift conversation. Play the role of a fictional, composite
donor — NOT any real person. Here's the made-up profile I want you to play: [INVENT A
PROFILE — "a retired teacher in her 70s, gives modestly each year, cares about literacy,
a bit skeptical of big asks"]. I'll practice opening a conversation and eventually making
an ask. Respond as this fictional donor would — including hesitation and hard questions.
Afterward, give me honest feedback on where I was warm, where I was pushy, and where I
missed a chance to listen.

Why this works A composite donor is the flight simulator — you crash safely and learn. Because the profile is invented, no real person's data touches the tool. And the after-action feedback is where the practice pays off: it names the pushy moment you couldn't feel in the moment.

Dignity Line

The donor here is fictional — that's the point. Never build the profile from a real prospect's actual details "to make it realistic." A made-up retired teacher teaches you just as well and costs a real person nothing.

23

Prepare questions for a discovery conversation

The task You want a first meeting to be about them, and you always default to talking about your org.

Help me prepare for a discovery conversation with a potential supporter — a first meeting
where my only job is to understand what they care about, not to pitch. Give me eight to
ten genuinely open, warm questions that help someone talk about their own values,
experiences, and what matters to them — without being intrusive or feeling like an
interview. Order them from easy and light to more meaningful, so the conversation can
warm up naturally.

Why this works Open questions, ordered from light to meaningful, is a real conversation architecture — it lets trust build instead of interrogating. And it fixes the fundraiser's most common reflex: filling the silence with your own mission instead of learning theirs. People remember what they say, not what you say.

24

Make sense of public research you gathered yourself

The task You've done the public research yourself. Now you have notes and want to prep smartly — without pasting the person in.

I've gathered some PUBLIC, non-sensitive information relevant to a potential funder's
interests (their published giving areas, their public mission, their stated priorities —
nothing private, no individual's personal data). Help me think through how our work might
genuinely connect to what they care about, and what respectful, relevant questions I could
ask to learn more. Here's the public information:
[PASTE ONLY PUBLIC, ORGANIZATIONAL INFORMATION — a foundation's published priorities, a
company's public giving focus — NEVER an individual's private details]

Why this works It draws the line right in the prompt: public and organizational, never private and individual. A foundation's published funding priorities are fair game — that's what they publish them for. A named individual's personal life is not. The tool helps you connect the dots you've already gathered legitimately.

Dignity Line

The safe version of "research" is organizational and public — a foundation's website, a company's giving page, an annual report. The unsafe version is a private individual's identity and history. Keep the first, never paste the second.

25

Draft outreach to a prospect (name added by you)

The task You're ready to reach out to a prospect. You want a warm first note.

Help me draft a short, warm first outreach email to a potential supporter I've identified
through public research. I don't want to pitch — I want to open a relationship. Use
placeholders [NAME] and [WHAT SEEMS TO CONNECT US — the public interest area]. About [100]
words. Warm, respectful, no ask in this first note. Position them as someone with something
valuable to offer, not as a target. End by inviting a conversation, low-pressure.

Why this works Placeholders keep the person out of the tool. "Position them as someone with something to offer, not as a target" flips the whole frame from extraction to relationship — which is both kinder and more effective. First contact with no ask is how you earn the second conversation.

Dignity Line

[NAME] stays a placeholder in the tool. You add the real name in your own email client, after. Same rule as the thank-you: the tool drafts the shape, you supply the human.

Data hygiene & admin

The unglamorous work that steals hours and never makes it into a job description. AI is quietly excellent here — as long as you keep the actual donor data out and use it on the structure of the work instead.

26

Design a data-entry standard for your CRM

The task Three people enter data three different ways and your database is a mess.

Our small team enters donor and gift information into our database inconsistently, and it's
creating a mess. Help me draft a simple, one-page "data entry standard" — clear rules for
how we record names, addresses, gift dates, gift types, and notes, so everyone does it the
same way. Keep it practical for a small team with no data specialist. Include a short
"why this matters" line for each rule so people actually follow it.

Why this works It builds you a standard, not a one-off fix. This is the single highest-leverage admin task in the sector — the CCNDR research names data management as the biggest skill gap in nonprofits, bigger than AI itself. Clean data is what makes everything else, including AI, actually work. The "why this matters" line is what gets your colleagues to follow it.

Dignity Line

Notice this prompt is about the rules, not the data. You never paste actual donor records to "help clean them up." You design the standard, then you apply it yourself inside your own system where the data lives safely.

27

Turn a process in your head into a written checklist

The task The gift-processing routine lives entirely in your head, and you're the only one who knows it.

I do a routine process at work that only lives in my head, and I need to write it down so
someone else could do it — or so I'm not the single point of failure. I'll describe the
process roughly; turn it into a clear, numbered step-by-step checklist, and flag any step
where something could go wrong so I can add a safeguard.

Here's the process, roughly:
[DESCRIBE THE PROCESS — the steps as you remember them, no specific donor data]

Why this works "Flag where something could go wrong" surfaces the risky steps you've stopped noticing because you do them on autopilot. Getting the process out of your head and onto paper is how a small shop stops being one sick day away from chaos.

28

Draft a policy conversation-starter for your ED

The task Your organization uses AI with no rules, and you're the one who knows it's a problem.

My organization uses AI tools informally with no policy, and I think we need at least a
basic one — especially around donor data. Help me draft a short, non-alarmist note to my
Executive Director suggesting we put some simple ground rules in place. Frame it as
protecting our donors and our reputation, not as a crisis. Keep it brief and constructive,
and suggest three or four basic rules a small nonprofit could actually adopt.

Why this works "Non-alarmist" and "not as a crisis" matter — you're raising this without sounding like you're accusing anyone. You're the canary who noticed the gap. This helps you raise it as a colleague protecting the organization, not a whistleblower. (When your ED is ready for the full version, Brazen makes an AI Policy Toolkit built for exactly this — but the conversation starts with a note, and this drafts it.)

Dignity Line

This is you being the person who protects the Dignity Line for your whole organization. Almost half of nonprofits have no AI policy at all (Virtuous, 2026). Being the one who raises it isn't overstepping — it's exactly the kind of quiet, responsible move that keeps donors safe.

29

Clean up messy text you already have

The task You've got a wall of text — pasted notes, a rough list — and you need it structured.

Below is some messy, unstructured text. Organize it into a clean, logical format —
[a table / a bulleted list / grouped categories — pick one]. Don't add any information I
didn't provide and don't invent anything. If something is ambiguous, flag it rather than
guessing. Just make what's here clear and usable.

[PASTE THE MESSY TEXT — with any personal donor/client data removed first]

Why this works "Don't add anything I didn't provide" keeps the tool honest — it's organizing, not embellishing. This is one of AI's genuinely best uses: taking chaos you already have and giving it a shape, fast, without changing the facts.

Dignity Line

Scrub personal donor and client details before pasting. If the messy text is a list of donors, this is exactly the kind of thing that stays in your own systems — clean it there, not here.

30

Build your own reusable prompt for a repeating task

The task You do the same kind of writing every week and you want a personal prompt you can reuse.

I do this task regularly: [DESCRIBE THE RECURRING TASK — e.g., "writing short program
updates for our email"]. Help me build a reusable prompt I can save and use every time,
with clearly marked [BRACKETS] for the parts that change each time. Make it specific
enough to give me consistent, good results, and include a reminder in the prompt itself
about keeping personal donor data out.

Why this works This is the prompt that makes the other twenty-nine yours. Instead of coming back to this library every time, you build your own small kit for the tasks you actually repeat. And asking it to bake the Dignity Line reminder right into your saved prompt means the safe habit travels with you. You stop borrowing time and start owning the skill — which was always the point.