Three documents run your year

The grant application, the donor letter, and the board report all land on the same few desks in a small nonprofit, and they eat the weeks your team should be spending on people. I rebuild how each one gets written: AI does the assembly, your staff keep the judgment.

Book The Honest Hour Start with the grant

Not an AI strategy.
Three rebuilt workflows

Nobody calls me because they want an AI strategy. They call because a grant is due Friday, the thank-you letters are three weeks behind, and the board packet still needs pulling together. So this page is organized around the documents themselves — what each one costs a small team, and what changes when the assembly work stops being manual.

AI drafts the assembly work Your team the judgment that stays human

The grant application

Most nonprofit grants aren't written by grant writers. They're written by program coordinators and executive directors already doing three other jobs — and every application starts from scratch. The organizational background, the program narrative, the outcomes, the budget justification. Again.

The rebuild keeps a living library of your organization's language, outcomes, and history, so a first draft takes an afternoon instead of a week. The library is yours, and your team runs it without me in the room.

The time ledger

Before
a week of evenings
After
one afternoon + your review

66.5% of Canadian charities hold restricted project funding — each grant with its own reporting requirements. Charity Insights Canada Project

Riverbend Neighbourhood House

Community resilience fund · renewal · draft 2

Sample · fictional org

Q2 · Describe your organization and the community you serve

“Riverbend has spent fourteen years pairing newcomer families in east Scarborough with trained neighbourhood volunteers — last year, 212 families across two sites, with program attendance holding above 90% for a third straight year.”

AI draftedassembled from the language library — the 2025 impact report and two past applications

“Demand has outgrown us: the waitlist sits at twenty families the waitlist reached 38 families in March, concentrated around the Warden Avenue site.”

Human editthe ED corrected the number — the tool can’t know what happened in March

“We’re asking for renewed support not to grow the model, but to stop turning families away from it.”

Voice checkread against past applications — it sounds like Riverbend, not like a bot

Reviewed & approved by the ED before submission

M. Osei · Executive Director

What a first draft actually looks like — and where people stay in it

The donor letter

Stewardship · first-gift thank-you

Composite

  • Opening

    AI drafts

    "Dear Priya — your first gift landed the same week our winter program filled up, and…"

    from gift record + segment

  • What your gift did

    AI drafts

    "Because of you, twelve more students had a tutor this term…"

    from your program notes

  • The line only you can write

    You add this

    "It was good to meet you at the open house — you asked the hard question about…"

    AI can't know the handshake

  • Sign-off & send

    You approve

    Signed by a person the donor has actually met.

    every send is approved by staff

In most small nonprofits, communications is one desk — newsletters, thank-yous, appeals, all of it. So donors who gave once don't hear anything real until the next ask, and most of them don't give twice.

AI drafts the stewardship pieces from your gift records and program notes. What it can't know — the conversation at the open house, the reason someone gave — your team adds in minutes instead of evenings. Nothing sends without a person behind it.

The time ledger

Before
three weeks behind
After
minutes per letter, same week

U.S. donor retention sits at 43.3%. For first-time donors it has stayed near 20% for years. AFP Fundraising Effectiveness Project

The board report

Finance lives in one system, donor records in another, program outcomes in a spreadsheet nobody's opened since March. Every quarter someone loses days pulling it all into one document and reconciling the numbers by hand.

AI synthesizes the raw exports into one structured draft. The part that matters — what the board should actually worry about this quarter — stays yours, because that's judgment, not assembly.

The time ledger

Before
days of copy-paste, every quarter
After
one working session + your framing

36% of U.S. nonprofits ended 2024 with an operating deficit, the highest in ten years of NFF survey data. Nonprofit Finance Fund

Board packet · Q2 · finance + programs

Composite

  • Financial summary

    AI drafts

    "Operating revenue is tracking four per cent ahead of budget; two grants land in Q3…"

    from your finance export

  • Program outcomes

    AI drafts

    "Attendance held at 92% across sites while the waitlist grew to 38 families…"

    from the tracker nobody's opened since March

  • What deserves the board's attention

    You frame this

    "Both of our largest grants renew in the same quarter next year. Here's the plan…"

    judgment, not assembly

  • Approve & circulate

    You approve

    Reviewed by the ED before it reaches a single board member.

    the packet goes out when you say it does

AI drafts. You decide

  1. Step 01

    AI generates

    First draft, template, or summary — ready in minutes instead of hours.

  2. Step 02

    You review

    Staff edit and add the organizational context only they have.

  3. Step 03

    You approve

    Nothing reaches a funder, donor, or board member without human sign-off.

And some work this never touches: crisis communications, case management decisions, therapeutic work, final funding decisions and high-stakes donor stewardship. That's your people's work — the point of clearing the documents is giving them the time to do it.

How I handle this responsibly

What executive directors ask me first

"Where does our donor data end up?"

In the systems you already control. The workflows we build run inside your own accounts, set up so your data isn't used to train anyone's model, and we start with the least sensitive documents first. Donor details only go into an AI tool with the donor's consent. The Responsible AI page covers the full setup.

"My board is skeptical. Honestly, so am I."

Good — skepticism is the right starting posture for this technology. Bring your board one rebuilt workflow with a visible human sign-off step, not an AI strategy deck. Approval authority stays with staff, so nothing about your governance changes.

"We don't have capacity to learn another system."

That constraint shapes everything here. Training happens inside your real work — we build the workflow on the grant you're actually submitting. The build itself takes about two afternoons; the rest of the two-week Implementation Package goes to training your staff so they can use AI well on their own. If it only works while I'm in the room, I've failed.

"Can we actually afford this?"

Every engagement has a fixed scope and a fixed price, published on the services page — no discovery-call surprise. Some funders treat this kind of work as capacity building, so it's worth asking yours.

Not sure which document to start with?

That's exactly what The Honest Hour is for. Book a free 60-minute conversation and we'll find the one eating the most time — or I'll tell you honestly that AI isn't your answer yet.

Book The Honest Hour Not ready to talk? Start with the writing.