It’s the Tuesday after the spring appeal. Gifts have been coming in since Friday, each one waiting for a thank-you. The development lead also has a volunteer rota with gaps in it, a board report due Thursday and a grant narrative that was due last week. The thank-yous are what slips.
Nonprofit teams are small for the size of the work. The same few people thank donors, recruit and schedule volunteers, run events, answer the board and write to funders. The mission work happens in whatever time is left.
Much of that admin is writing and looking things up, which is what an AI assistant does well. Some of it is trust: a donor’s trust that their gift was noticed, a funder’s trust that the numbers are real, a beneficiary’s trust that their story is safe. That part stays with people. Here’s how we’d draw the line.
What AI should do
Donor thank-yous
Every gift deserves a thank-you, and the first gift deserves a good one. In practice the notes pile up, go out late, or become one template with a name swapped in. An assistant can draft a note for each first and returning gift, in your voice, using what you know: which appeal they gave to, whether they volunteer, whether this is their first gift in years. A person reads each one, edits, and sends.
The donor gets a note that sounds like you and arrives while the gift is still fresh. Your development lead spends their time on the handful of notes that need their own words.
Stopped gifts
A recurring gift can stop for very different reasons. Often the card expired. Sometimes the donor lost a job, moved on, or was unhappy about something. Those need completely different responses, and a list of “lapsed” donors doesn’t tell them apart.
An assistant can sort them into “card expired” and “worth a call”, with the reason for each. A card update note is a good candidate for automatic sending once you trust it. A donor worth a call gets a call, from a person.
Volunteers
Volunteers are often your most committed supporters, and the easiest to wear out. An assistant can suggest rotas that spread the load, flag the people who haven’t had a break, and set up an event and its sign-up from a sentence. After the event it can tell you who came for the first time, who came back, and who didn’t show.
Reports
Board reports and grant reports take a lot of hours for work that is mostly gathering. An assistant can draft the month in plain English from your real numbers, with every figure linked to where it came from. For grant reports, it can draft the narrative from participation and outcomes on your record.
Give
- A thank-you drafted for every first and returning giftCommunity
- Stopped gifts sorted into “card expired” and “worth a call”Community
Serve and gather
- Rota suggestions that spread the load and flag burnoutCommunity
- Events, sign-ups and reminders set up from a sentenceCommunity
- An after-event summary: first-timers, returners, no-showsCommunity
Report
- Board report drafted from your real numbers, every figure sourcedCommunity
- Summaries for each chapter, campus or group leaderCommunity
- Grant report narrative for nonprofitsAdd-on
What we build, and which tier each part is in.
What AI shouldn’t do
Invent a number or an outcome
A report to a funder is a statement of fact. If an assistant writes “participants reported higher confidence” and nobody measured it, your organization has now said something untrue. Every figure in a drafted report should link to where it came from, and every claim about outcomes should be checked by the person who signs the report. If a figure has no source, cut it.
Write to beneficiaries about sensitive matters on its own
Many nonprofits serve people at hard moments: housing, health, immigration, recovery. Messages about those things need a person who knows the situation. An event reminder can go automatically. A note to someone in crisis cannot.
Decide who is worth your time
Sorting stopped gifts is useful. Ranking donors by a score and letting the bottom of the list go quiet is not. Show the reason behind every flag and let a person decide. There’s more on that in why a reason beats a score.
Ignore what people asked for
Some donors don’t want phone calls. Some volunteers only want texts. Whatever tool you use, check that it honors contact preferences before any draft reaches the approval queue, not after.
What it looks like in practice
A thank-you for every gift
First and returning gifts each get a note drafted in your voice, and stopped gifts are sorted into “card expired” and “worth a call”.
The grant report, drafted
Participation and outcomes pulled from your own records into a narrative, with every figure sourced. A scoped add-on.
All of this depends on donors, volunteers and participants being the same person across your tools. If giving, events and volunteering each keep their own list, no assistant can tell you that your most loyal volunteer just stopped giving. One member record is what makes it possible.
Where to start
Start with thank-yous. They are frequent, the risk is low, and everyone on the team can see whether the drafts are good. Run them for a month with a person approving each one, and keep a note of what gets changed. Then move to stopped gifts, then reports.
Before any of it, write down which messages must always be read by a person. Share it with your board. It’s a short document and it will settle a lot of later questions.
What we build
We build an assistant into the nonprofit platforms we make, with everyone’s whole story in one place your organization owns. Your assistant runs in your cloud, on your own data, under your own account. It sees only what the person asking may see, never trains anyone’s model on your members, and drafts rather than sends.
AI usage is paid to your own provider at cost, like hosting, with no markup and no platform fee. We estimate it in your scope. Grant report narratives are a scoped add-on; thank-yous, stopped-gift sorting, rotas and board reports are in the Community tier. The AI Assistant page has the full list and every commitment it keeps.
Questions we hear
Will AI thank-you notes sound generic?
They shouldn’t, if the draft uses what you know about the donor and is written in your voice, and a person reads it before it goes. The assistant we build drafts in your organization’s voice, learned from your past posts, and a person on your team approves every message.
Can the assistant write our grant report?
It can draft the narrative from participation and outcomes you’ve recorded, with every figure sourced, as a scoped add-on. Your team checks every figure and every claim before it goes to a funder. AI usage is paid to your own provider at cost, like hosting, with no markup and no platform fee. We estimate it in your scope.
Does it send messages to donors automatically?
Every message to a member waits for a person to approve it. You can switch on automatic sending for low-risk jobs you choose, like event reminders. Anything pastoral is never automatic.
Is our donor data used to train an AI model?
No. Your provider account runs under business terms that exclude training on your data.