Skip to content

AI prompts for nonprofit fundraising, sorted by the job.

68 things to ask an AI about prospect research, major gifts, donor retention, donor data, web pages and appeals. Each one says what you get back and what a person still decides, and 33 also work in a general AI assistant.

Ask the Prospect Research agent

Should the Thompsons move into major gift qualification?

How to use these prompts

  1. Pick the version that fits your tool

    The short ask is what you type to an Ignite agent, which already reads your CRM. The longer prompt works in a general AI assistant: fill in the [brackets] with your own facts.

  2. Keep donor data out of public tools

    Remove names and contact details before you paste anything into a general AI assistant, or use the tool your organization approved. The free AI policy template helps you set the rules.

  3. A person checks every answer

    Each card names the step that stays with you. Treat the output as a draft: check the evidence, then decide.

Prospect research prompts

Prospect research prompts help a researcher decide who to qualify next and what to verify by hand first. In Ignite, the Prospect Research agent reads the CRM record and wealth screening, and a write to a prospect record waits at the approval gates set for your organization.

Meet the Prospect Research agent

Qualify a prospect

Changes no records

Ask the Prospect Research agent

“Should the Thompsons move into major gift qualification?”

You get back
A recommendation with the evidence used, confidence, blockers, a next step and items to verify by hand.
Your step
You decide whether the prospect moves into qualification.
In a general AI assistant

You are helping a nonprofit prospect researcher. Here is what we know about one prospect, with names and contact details removed: [giving history: amounts and dates], [engagement: events, volunteering, board or committee roles], [wealth indicators we already hold, with the source of each]. Should this prospect move into major gift qualification? Answer yes, no or not yet. Give the evidence for your answer, what would change it, and three things I should verify by hand before I decide.

Explain a score

Changes no records

Ask the Prospect Research agent

“Why did this prospect score high on affinity?”

You get back
The scored dimensions and the evidence behind each one.
Your step
You check the evidence before you rely on the score.

Ignite only: this job needs your live CRM data.

Check a prospect’s wealth signals

Changes no records

Ask the Prospect Research agent

“What wealth signals do we have for Maria Alvarez?”

You get back
Wealth screening signals, kept apart from what the CRM record says.
Your step
You confirm the match is the right person.

Ignite only: this job needs wealth screening data.

Find the warmest path

Changes no records

Ask the Prospect Research agent

“What is our warmest path to Robert Kim?”

You get back
The strongest relationship signals, the likely warm path, the missing data and an outreach path.
Your step
You choose who makes the introduction.
In a general AI assistant

You are helping a nonprofit fundraiser plan an introduction. Here are the documented connections between our organization and one prospect, with names replaced by roles: [for example: a board member who serves with them on another board], [a volunteer who works at the same company], [a staff member who met them at an event]. Which path is warmest, and why? List what is missing from this picture, then suggest a first outreach step for the strongest path.

Find individual prospects

Changes no records

Ask the Prospect Research agent

“Find major gift prospects in our donor base from giving and engagement history.”

You get back
Five to ten candidates, each with a score, a recommendation, an ask band, identity confidence and the strongest evidence.
Your step
You choose which candidates to research further.

Ignite only: this job needs your live CRM data.

Find corporate prospects

Changes no records

Ask the Prospect Research agent

“Find corporate prospects from our sponsorship and workplace giving signals.”

You get back
Five to ten organizations with their giving, a score, a recommendation, an ask band and key contacts.
Your step
You pick which companies to approach.

Ignite only: this job needs your live CRM data.

Compare prospects

Changes no records

Ask the Prospect Research agent

“Compare these three prospects and tell me which one to qualify first.”

You get back
A short comparison with the evidence for each prospect.
Your step
You set the research order.
In a general AI assistant

Compare three prospects for a nonprofit major gift program. Names and contact details are removed. Prospect A: [giving, engagement, connections to us]. Prospect B: [the same facts]. Prospect C: [the same facts]. Which one should we qualify first? Put the comparison in a short table with capacity signals, affinity signals, connection to us and gaps. Then give your pick and the reason in two sentences.

Check the ask band

Changes no records

Ask the Prospect Research agent

“What ask band does the score give Maria Alvarez, and what holds it back?”

You get back
The ask band from the platform score, with the reasons that hold it back.
Your step
You decide whether the band fits what you know.

Ignite only: this job needs your live CRM data.

Keep the research

Saves after your approval

Ask the Prospect Research agent

“Add this research to the prospect’s profile.”

You get back
The exact change the agent proposes, held until a person approves it.
Your step
You approve or reject the change.

Ignite only: this job needs your live CRM data.

Major gift prompts

Major gift prompts help a gift officer prepare for each stage of a major gift: discovery, cultivation and the ask. In Ignite, the Major Gift agent reads the relationship history and recommends a move, and the gift officer decides the move and makes the ask.

Meet the Major Gift agent

Prepare for a donor meeting

Changes no records

Ask the Major Gift agent

“Prepare a brief for my lunch with Sarah Williams tomorrow.”

You get back
Why the meeting matters, likely motivations, talking points, questions, risks and the next move.
Your step
You lead the conversation.
In a general AI assistant

I am a major gift officer at a nonprofit. Tomorrow I meet a donor for [purpose: discovery, cultivation or an ask conversation]. Here is the relationship so far, with names and contact details removed: [giving history], [programs they care about], [past conversations and what they said]. Write a one-page brief: why this meeting matters, likely motivations, three talking points, three questions to ask, sensitivities to avoid and the next move to aim for.

Plan a discovery visit

Changes no records

Ask the Major Gift agent

“What should I explore in a discovery visit with Dana Ortiz?”

You get back
Questions to explore, likely motivations and sensitivities.
Your step
You run the visit and record what you learn.
In a general AI assistant

I have a discovery visit with a donor. It is our first deep conversation, and I will not make an ask. What we know, with names and contact details removed: [giving history], [how they came to us], [programs they have taken part in]. Give me eight open questions that help me learn what they care about, the motivations to listen for and the topics to handle with care.

Plan cultivation

Changes no records

Ask the Major Gift agent

“Build a 90-day cultivation plan for this prospect.”

You get back
Steps for days 30, 60 and 90, the risks and the first priority.
Your step
You choose which steps to take.
In a general AI assistant

Build a 90-day cultivation plan for a nonprofit major gift prospect. What we know, with names and contact details removed: [relationship history], [interests], [last contact and what came of it], [our goal for this relationship]. Give steps for the first 30, 60 and 90 days, the risk at each stage and the one thing to do first. Keep to moves a gift officer can make in person or by phone.

Choose the next move

Changes no records

Ask the Major Gift agent

“What’s the best next move with the Chens?”

You get back
One move, why it wins, what blocks it and what would change it.
Your step
You choose the move and make it.
In a general AI assistant

Here is where a major gift relationship stands, with names and contact details removed: [stage: discovery, cultivation or ask], [the last three contacts and what came of each], [open questions]. Recommend one next move. Say why it beats the other options, what could block it and what new information would change your recommendation.

Frame the ask

Changes no records

Ask the Major Gift agent

“How should we frame the ask for the Peterson family?”

You get back
The best current ask hypothesis, the evidence, framing, timing and items to verify.
Your step
You set the amount and make the ask.
In a general AI assistant

Help me frame a major gift ask. Donor context, with names and contact details removed: [giving history and largest gifts], [programs they fund or ask about], [what they have said about their goals], [the project and what it needs]. Suggest how to frame the ask: the purpose to lead with, a range to consider and why, when to ask and what to confirm before the conversation. Label anything you infer as a hypothesis.

Test an ask amount

Changes no records

Ask the Major Gift agent

“Is $50,000 the right ask for the Peterson family for the capital campaign, or is it too early?”

You get back
An ask range, or a reason to stay in cultivation longer.
Your step
You set the amount and the timing.
In a general AI assistant

Is [amount] the right ask for this donor for [campaign or project], or is it too early? Context, with names and contact details removed: [giving history], [engagement over the last two years], [what they have said about the project]. Answer with one of two calls, ask now or keep cultivating. Then give a range that fits the evidence and the two facts that matter most to your answer.

Check risks before the ask

Changes no records

Ask the Major Gift agent

“What risks or sensitivities should I know before I ask the Chens?”

You get back
Risks, sensitivities and items to verify by hand.
Your step
You decide whether to go ahead.
In a general AI assistant

Before I make a major gift ask, help me check for risks. Context, with names and contact details removed: [recent changes in giving], [notes from the last few conversations], [concerns they raised, in general terms], [other asks they had from us this year]. List the risks and sensitivities you see, how serious each one is and what I should verify before the ask.

Explain the AI probability

Changes no records

Ask the Major Gift agent

“Explain the AI probability on this opportunity.”

You get back
Why the AI probability differs from the stage probability, with the evidence.
Your step
You decide how much weight to give it.

Ignite only: this job needs your live CRM data.

Balance gift officer portfolios

Changes no records

Ask the Major Gift agent

“Rebalance the portfolios across our gift officers.”

You get back
Where workloads are uneven across gift officers, based on load, win rate and stale opportunities, with reassignment ideas for the opportunities you name.
Your step
You make any reassignment. The agent never assigns.

Ignite only: this job needs your live CRM data.

See what closed gifts had in common

Changes no records

Ask the Major Gift agent

“What cultivation moves show up most in the gifts we closed?”

You get back
Activity sequences that appear more often in won opportunities than in lost ones. Treat it as a pattern to test.
Your step
You decide what to try in your own portfolio.

Ignite only: this job needs your live CRM data.

Donor retention prompts

Donor retention prompts help an annual giving or donor services team find the donors who need attention and the best move to keep or win back each gift. In Ignite, the Retention agent reads giving and payment history and recommends a move; a person makes the contact, and the agent does not change billing.

Meet the Retention agent

Find recurring donors at risk of lapsing

Changes no records

Ask the Retention agent

“Which recurring donors are at risk of lapsing this month?”

You get back
A risk view with the evidence for and against, and items to verify.
Your step
You choose who to contact.
In a general AI assistant

Help me read lapse risk for recurring donors. For each donor, with names and contact details removed, I list [gift amount and schedule], [months on the plan], [recent payment problems] and [recent engagement, such as email opens or event attendance]. Sort them by lapse risk. For each one, give the signals that raise concern, the signals that calm it and one thing to check before we reach out.

Rank the week’s calls

Changes no records

Ask the Retention agent

“Which recurring donors need attention first this week?”

You get back
A ranked shortlist with the filters used, a reason for each donor and a recommended move.
Your step
You make the calls.

Ignite only: this job needs your live CRM data.

Recover one recurring gift

Changes no records

Ask the Retention agent

“This donor’s recurring gift stopped in March. What is the best recovery move?”

You get back
What the record shows, the best recovery move and the caveats.
Your step
You make the contact.
In a general AI assistant

A donor’s recurring gift stopped in [month]. What I know, with the name and contact details removed: [plan amount and schedule], [how long they gave], [any payment errors we can see], [other contact since]. What is the best recovery move? Do not guess at why the payment stopped. Tell me what to confirm first, the move to make and the message to lead with.

Choose a save action

Changes no records

Ask the Retention agent

“What’s the best save action for this donor?”

You get back
One action, why it wins, what blocks it and what would change the call.
Your step
You take the action.
In a general AI assistant

A recurring donor may be about to stop giving. Here is the picture, with the name and contact details removed: [plan amount and schedule], [what changed: a failed payment, lower engagement, a complaint], [their history with us]. Recommend one save action: outreach, a stewardship touch, investigate first or hold. Say why it wins, what could block it and what would change your answer.

Read one donor’s retention picture

Changes no records

Ask the Retention agent

“Summarize the retention picture for Marguerite Osei.”

You get back
The retention picture, the strongest signals, the unknowns and the next action.
Your step
You decide what happens next.
In a general AI assistant

Summarize the retention picture for one donor, with the name and contact details removed: [giving history for the last three years], [recurring plan status], [engagement], [last contact]. Give the overall picture in two sentences, the strongest signals for and against, what we do not know and the next action.

Find upgrade candidates

Changes no records

Ask the Retention agent

“Who are our best candidates for an upgrade from $25 to $50 a month?”

You get back
Stable donors with loyalty signals, the caveats and an upgrade idea for each.
Your step
You decide who gets the ask.
In a general AI assistant

Which of these recurring donors are ready to be asked to increase their gift? Names and contact details are removed: [for each donor: current amount, months on the plan, payment history, other gifts, engagement]. Pick the strongest candidates and explain why. Flag any donor who looks fragile and should get a thank-you instead, and suggest an upgrade amount for each pick.

Find reactivation candidates

Changes no records

Ask the Retention agent

“Find lapsed donors from last year who are good reactivation candidates.”

You get back
A ranked shortlist with evidence, caveats and a next move for each.
Your step
You choose who to reach out to.
In a general AI assistant

Here are donors who gave in the past but not this year, with names and contact details removed: [for each: last gift date and amount, years of giving, how they gave, engagement since]. Rank the ten best reactivation candidates. For each one, give the evidence of past loyalty, a caveat and the first move to try.

Find last year’s donors who have not given

Changes no records

Ask the Retention agent

“Who gave last fiscal year but has not given this year?”

You get back
Donors who gave last fiscal year but not this one, by your organization’s own definition.
Your step
You decide how to reach them.
In a general AI assistant

Here are donors who gave last fiscal year and have not given this year, with names and contact details removed: [for each: last gift date, amount, years given, channel]. Group them for outreach, for example long-time donors, first-year donors and event-only donors. Say which group to contact first and why.

Recover workplace donors

Changes no records

Ask the Retention agent

“Which workplace donors pledged last year but have not pledged this year?”

You get back
A ranked shortlist of workplace donors, with caveats.
Your step
You plan the outreach.

Ignite only: this job needs your live CRM data.

Work a saved segment

Changes no records

Ask the Retention agent

“From our lapsed-donor segment, who should we call first?”

You get back
A ranked shortlist from the saved segment.
Your step
You make the calls.

Ignite only: this job needs your live CRM data.

Donor data questions

Anyone on the team can ask these donor data questions in plain words, without a report request. In Ignite, the Constituent Support agent answers from live CRM data with its sources shown, and it does not change donor, gift or household records.

Meet the Constituent Support agent

Get the full picture on a donor

Changes no records

Ask the Constituent Support agent

“Give me the full picture on Marguerite Osei.”

You get back
Giving, engagement, household and context in one answer.
Your step
You use it to prepare your next contact.

Ignite only: this job needs your live CRM data.

Catch up on a donor

Changes no records

Ask the Constituent Support agent

“What is new with the Alvarez household since our last contact?”

You get back
Recent activity, notes, interactions and insights since that contact.
Your step
You decide whether to reach out.

Ignite only: this job needs your live CRM data.

Find people from their notes

Changes no records

Ask the Constituent Support agent

“Who’s mentioned an interest in golf?”

You get back
The people whose notes match, with the note.
Your step
You decide who to invite or follow up with.

Ignite only: this job needs your live CRM data.

Total a campaign

Changes no records

Ask the Constituent Support agent

“What’s the total giving for the Summer 2025 campaign?”

You get back
The total and the gift count, with the date range stated.
Your step
You check the date range before you share the number.

Ignite only: this job needs your live CRM data.

Read a KPI

Changes no records

Ask the Constituent Support agent

“What’s our donor retention rate this year?”

You get back
The rate, explained in plain words.
Your step
You decide what the number means for your plan.

Ignite only: this job needs your live CRM data.

Compare two fiscal years

Changes no records

Ask the Constituent Support agent

“How does giving this fiscal year compare with last year?”

You get back
Totals and the trend, with both periods stated.
Your step
You take the questions to your team.
In a general AI assistant

Here are our gift totals by month for this fiscal year and last fiscal year: [a table of month and total for each year]. Our fiscal year starts in [month]. Compare the two years to date, point out the months that explain most of the difference and list two questions I should ask our team about the trend.

Rank donors by lifetime giving

Changes no records

Ask the Constituent Support agent

“Show me our top 10 donors by lifetime giving.”

You get back
A ranked table, with a link to any record you ask for.
Your step
You decide who needs a stewardship touch.
In a general AI assistant

Here is a giving export with names replaced by donor IDs: [donor ID, lifetime giving, first gift date, last gift date]. Rank the top 10 by lifetime giving in a table. Then note any donor in the top 10 whose last gift is more than 18 months old.

Check open pledges

Changes no records

Ask the Constituent Support agent

“What pledges does Robert Kim have open, and what is still due?”

You get back
The open pledges, with their balances.
Your step
You plan the reminder or the thank-you.

Ignite only: this job needs your live CRM data.

Read a household

Changes no records

Ask the Constituent Support agent

“Who is in the Thompson household, and who is the head of household?”

You get back
The household members, with the head of household marked.
Your step
You choose how to address the household.

Ignite only: this job needs your live CRM data.

Check an employer

Changes no records

Ask the Constituent Support agent

“Where does Maria Alvarez work, and who is the primary contact there?”

You get back
The employer, the job title and the organization’s primary contact.
Your step
You decide whether a workplace or matching-gift conversation fits.

Ignite only: this job needs your live CRM data.

Donor enrichment prompts

Donor enrichment prompts help a researcher add wealth screening to what the CRM already knows about a donor, and keep the two apart. In Ignite, the Donor Enrichment agent uses WealthEngine data and does not change donor, gift or household records; a person confirms the match and decides the outreach.

Meet the Donor Enrichment agent

Enrich one donor

Changes no records

Ask the Donor Enrichment agent

“Enrich this donor and summarize their giving potential.”

You get back
The CRM profile, an enrichment summary, key wealth signals, an outreach idea and notes on data quality.
Your step
You confirm the match before you act on it.

Ignite only: this job needs wealth screening data.

Separate fact from inference

Changes no records

Ask the Donor Enrichment agent

“Enrich Dana Ortiz and separate what WealthEngine reported from what you inferred.”

You get back
Source facts and inferred recommendations, kept apart.
Your step
You decide how much weight the inference gets.

Ignite only: this job needs wealth screening data.

See wealth signals beside the CRM record

Changes no records

Ask the Donor Enrichment agent

“What wealth signals do we have for Michael Park?”

You get back
Key wealth signals next to the CRM record.
Your step
You confirm the signals belong to the right person.

Ignite only: this job needs wealth screening data.

Check match quality

Changes no records

Ask the Donor Enrichment agent

“How strong is the match on this enrichment?”

You get back
The match quality, with weak matches flagged.
Your step
You set weak matches aside until you confirm them.

Ignite only: this job needs wealth screening data.

Enrich by email address

Changes no records

Ask the Donor Enrichment agent

“Enrich the profile for donor@example.com.”

You get back
The matched record, confirmed before enrichment runs.
Your step
You check that the record is the right donor.

Ignite only: this job needs wealth screening data.

Turn a profile into outreach

Changes no records

Ask the Donor Enrichment agent

“What outreach would you suggest based on this profile?”

You get back
A suggested first move, labeled as inference.
Your step
You decide the outreach.
In a general AI assistant

Here is a donor profile, with the name and contact details removed: [giving history], [interests], [wealth indicators we hold, with the source of each]. Suggest a first outreach move. Keep what the facts show apart from what you infer, and say what I should confirm before I act.

Screen a list

Changes no records

Ask the Donor Enrichment agent

“Run a batch enrichment for these 100 prospects and track status.”

You get back
A submitted batch and its status.
Your step
You choose the records before the batch goes to screening.

Ignite only: this job needs wealth screening data.

Rank a screened batch

Changes no records

Ask the Donor Enrichment agent

“Which records in this batch have the strongest major gift indicators?”

You get back
The strongest records, and why.
Your step
You choose who moves on to research.

Ignite only: this job needs wealth screening data.

Check on a batch

Changes no records

Ask the Donor Enrichment agent

“Where is my batch enrichment?”

You get back
The batch status.
Your step
You plan the review when it finishes.

Ignite only: this job needs wealth screening data.

Web page prompts

Web page prompts help a team draft donation, event, volunteer, signup and survey pages without starting from a blank page. In Ignite, the Web Designer agent drafts the page in Ignite Design Studio, and a person reviews and publishes it.

Meet the Web Designer agent

Draft a volunteer page

Draft for your review

Ask the Web Designer agent

“Draft a volunteer page for the Riverside Cleanup.”

You get back
A draft page built around the volunteer form.
Your step
You review it and publish it.
In a general AI assistant

Write the copy for a volunteer signup page for [event or program]. Details: [date, place, what volunteers will do, who it helps, what to bring]. Give me a headline, a two-sentence intro, three short sections (what you will do, who it helps, what to know before you come) and the text for the signup button. Use plain words.

Draft an event registration page

Draft for your review

Ask the Web Designer agent

“Draft a registration page for the spring gala.”

You get back
A draft page built around the registration form of an event you set up in Ignite.
Your step
You review it and publish it.
In a general AI assistant

Write the copy for a registration page for [event]. Details: [date, place, ticket types, what the money supports, dress code]. Give a headline, a short intro, a schedule section, a section on what each ticket supports and the button text.

Draft a signup page

Draft for your review

Ask the Web Designer agent

“Draft a signup page for updates on the shelter expansion.”

You get back
A draft page built around the signup form.
Your step
You review it and publish it.
In a general AI assistant

Write the copy for a page where people sign up for updates on [project]. Details: [what the project is, why it matters, how often we will send updates]. Give a headline, a short intro, three reasons to sign up and the button text. Do not promise an update schedule I have not given you.

Draft a survey page

Draft for your review

Ask the Web Designer agent

“Draft a short survey page for gala guests.”

You get back
A draft survey page.
Your step
You review it and publish it.
In a general AI assistant

Write a short survey for guests who attended [event]. We want to learn [for example: what they enjoyed, what to change, whether they would come back]. Give a one-sentence intro and no more than six questions, with a mix of rating and open questions. Keep it under two minutes to finish.

Pick a direction for a giving page

Draft for your review

Ask the Web Designer agent

“Suggest directions for a year-end giving page for recurring donors.”

You get back
Design directions, with a recommended fit.
Your step
You pick the direction.
In a general AI assistant

I am planning a year-end giving page for donors who already give on a recurring plan. Give me three different directions for the page. For each one: the headline, the main message, what the page shows first and the kind of photo that fits. Then say which direction you would pick and why.

Make a small edit

Draft for your review

Ask the Web Designer agent

“Shorten the headline and use a photo with families in it.”

You get back
The edit, and a note on what changed.
Your step
You accept the edit or ask for another.
In a general AI assistant

Here are the headline and intro from our donation page: [paste the text]. Shorten the headline to eight words or fewer and keep the meaning. Give me three options, and describe the kind of photo that fits each one.

Set campaign colors

Draft for your review

Ask the Web Designer agent

“Give the harvest festival page warm autumn colors.”

You get back
A palette that fits the page, checked for contrast.
Your step
You approve the look.

Ignite only: this job works inside Ignite Design Studio.

Redesign a page

Draft for your review

Ask the Web Designer agent

“Redesign the general donation page to feel warmer.”

You get back
A rebuilt layout, with the form connections kept.
Your step
You review it and publish it.

Ignite only: this job works inside Ignite Design Studio.

Check the finished page

Changes no records

Ask the Web Designer agent

“Why does the hero image crop badly on phones?”

You get back
Layout, color and image problems found on the rendered page.
Your step
You choose which fixes to make.

Ignite only: this job works inside Ignite Design Studio.

Work through reviewer comments

Draft for your review

Ask the Web Designer agent

“Work through the reviewer comments on this page.”

You get back
Edits tied to each comment, with each thread marked as addressed.
Your step
You check the edits before you publish.

Ignite only: this job works inside Ignite Design Studio.

Appeal and email prompts

Appeal and email prompts help a marketing team write, tighten and personalize appeals, then read how they did. In Ignite, the Marketing Agent drafts and edits in Ignite Design Studio, and a person sends, schedules and issues every piece.

Meet the Marketing Agent

Write subject lines

Changes no records

Ask the Marketing Agent

“Give me subject line options for the year-end appeal.”

You get back
Subject lines and preheaders to choose from.
Your step
You pick the line.
In a general AI assistant

Write ten subject lines and matching preheaders for our year-end appeal email. The appeal is about [the need, and what a gift makes possible]. Mix short and specific lines. No false urgency, no all caps, and no “Re:” or “Fwd:”.

Tighten an appeal

Draft for your review

Ask the Marketing Agent

“Tighten the opening and make the gift button the only call to action.”

You get back
The edited draft in Ignite Design Studio.
Your step
You review it and send it.
In a general AI assistant

Here is the draft of our fundraising email: [paste the text]. Tighten the opening so the reader knows in two sentences why we are writing now. Remove every call to action except the gift button. Keep our facts and numbers as they are, and show me what you cut.

Personalize an appeal

Draft for your review

Ask the Marketing Agent

“Add the donor’s first name and a suggested ask to this appeal.”

You get back
The draft, with personalization tokens in place.
Your step
You preview it and send it.

Ignite only: this job works inside Ignite Design Studio.

Check who an email will reach

Changes no records

Ask the Marketing Agent

“Who is in the lapsed-donor audience for this email?”

You get back
Who the email will reach.
Your step
You confirm the audience before you send.

Ignite only: this job needs your live CRM data.

Tighten a story

Draft for your review

Ask the Marketing Agent

“Tighten this story about the food pantry expansion.”

You get back
The tightened text, to accept or discard.
Your step
You accept it or keep your version.
In a general AI assistant

Tighten this story for our donor update: [paste the story]. Keep it under 250 words and keep every fact and quote. Lead with the person or the moment, and end with what a donor’s gift made possible.

Draft a story from notes

Draft for your review

Ask the Marketing Agent

“Draft a story for this issue from my notes on the food pantry expansion.”

You get back
One draft story, to accept or discard.
Your step
You edit it and decide whether it runs.
In a general AI assistant

Draft a 200-word story for our donor update from these notes: [paste your notes]. Lead with one person or moment, use only the facts in my notes and end with what donors made possible. Flag anything you had to assume.

Rewrite a gift acknowledgement

Draft for your review

Ask the Marketing Agent

“Rewrite our gift acknowledgement so the thank-you comes first. Keep the receipt language.”

You get back
The rewritten template, with the compliance content kept.
Your step
You work on a copy and review it before you make it active.
In a general AI assistant

Here is our gift acknowledgement letter: [paste the letter template, with no donor name or address]. Rewrite it so the thank-you comes first and reads as personal. Do not change the receipt or tax language, and mark it so I can see you left it as written.

Redesign a year-end statement

Draft for your review

Ask the Marketing Agent

“Make our year-end statement easier to read on a phone.”

You get back
The redesigned template, with the receipt and compliance content kept.
Your step
You work on a copy and review it before you make it active.

Ignite only: this job works inside Ignite Design Studio.

Read appeal results

Changes no records

Ask the Marketing Agent

“How did the spring appeal do?”

You get back
The appeal’s results and revenue, explained.
Your step
You decide what to change next time.
In a general AI assistant

Here are the results of our spring appeal: [sent, opened, clicked, gifts and revenue, by channel if you have it]. Our goal was [goal]. Explain how it did in plain words, what stands out and two things to test in the next appeal.

Compare appeals

Changes no records

Ask the Marketing Agent

“Compare the results of our last three appeals.”

You get back
The results side by side, explained.
Your step
You decide what to repeat.
In a general AI assistant

Compare our last three appeals: [for each: date, audience size, channel, opens, clicks, gifts, revenue]. Put them in a table, explain the differences that matter and say what the best one did that the others did not.

What makes a good fundraising prompt

A good prompt has four parts. A weak answer often comes from a prompt that leaves one out.

Vague

“Tell me about our donors.”

Specific

“Which recurring donors are at risk of lapsing this month? Rank them and give the evidence for each one.”

  1. The record

    Who or what the question is about: one donor, a segment, a campaign, a page.

  2. The question

    One decision you need to make, asked the way you would ask a colleague.

  3. The time window

    This month, last fiscal year, since the last contact. Without one, the answer guesses.

  4. The output

    A ranked list, a one-page brief, a table, three options. Say which one you want.

Why the Ignite ask is shorter

A general AI assistant knows only what you paste into it, so the prompt has to carry the record, the time window and the output you want. An Ignite agent starts from your live CRM data, so one line is enough.

It reads within your permissions

The agent reads within the permissions your organization already sets. You do not paste donor data anywhere to get the answer.

It shows where the answer came from

Answers from your CRM show the evidence behind them: the records, filters and date ranges used. You can check the work before you act on it.

Changes wait for a person

Any change an agent can make follows the approval gates set for your organization. Drafts wait for a person to publish or send them.

How Ignite governs its agents

Questions about AI prompts for fundraising

For the basics, the leader’s guide to AI agents explains how agents differ from chatbots and where approval fits.

Can I use these prompts in ChatGPT, Claude or Copilot?

Yes. The longer prompt on each card is written for a general AI assistant. Fill in the brackets with your own facts, and remove names and contact details first. Cards marked “Ignite only” need live CRM data, wealth screening or Ignite Design Studio, which a general assistant does not have.

Is it safe to put donor data in an AI prompt?

Only in a tool your organization approved for donor data. In a public AI tool, remove names, contact details and anything else that identifies a person. If your team has no policy yet, the free AI policy template is a place to start. The general prompts on this page are written to work without names.

Why are some prompts marked Ignite only?

Some jobs need data a general assistant does not have: live giving and payment history, platform scores or wealth screening. For those, the card gives the Ignite ask only, so you do not get a confident answer built on nothing. On this page that covers most of the donor data and donor enrichment cards.

Do Ignite agents need a long prompt to work?

No. You ask in plain words, the way you would ask a colleague, and the agent starts from your live CRM data. Agents work when someone on your team asks. A general AI assistant needs more, because it knows only what you paste in: the record, the question, the time window and the output you want.

Who checks what the agent writes?

A person. Each card names the step that stays with you. Changes follow the approval gates set for your organization, drafts wait for someone to publish or send them, and an answer on its own changes no records. Check the evidence behind an answer before you act on it, as you would with a first draft from a colleague.

Bring the questions your team asks most.

Book a working session and see an agent answer them from a live record, evidence and all.