Most nonprofits arrive at the AI question with the same gap: they have seen headlines about time savings and efficiency gains, but they have no map for which tasks in their specific org are actually worth trying. The use cases in those headlines are either too vague (“AI can help with communications”) or too advanced (“AI can predict donor lapse”) to connect to the reality of a six-person team running on a tight budget. This piece gives five specific answers, with honest expectations for each. Not every use case fits every org, and the one that looks most compelling to your board may not be the right starting point for your team.
What makes a task a good fit for AI?
AI works well on tasks that are repetitive, text-based, and low-stakes if something goes wrong. The five use cases below share those three qualities.
The filter is simple. Good AI tasks involve a lot of routine writing or summarising, the output needs human review before it goes anywhere important, and a mistake costs time to fix rather than causing harm to a person or relationship. If a task clears all three, AI is worth trying. If any one fails, it deserves a different question.
This also means there is a list of things kept deliberately off the five. Tasks involving sensitive beneficiary information, crisis communications, and individual legal or HR matters fall into a different category, covered at the end of this piece. The three-part filter is something you can apply yourself when you wonder whether AI would help with a task your team is stuck on.
Drafting grant proposals
AI can produce a working first draft of a grant proposal in a fraction of the time it takes to write one from scratch, but it cannot replace the strategic thinking that wins the grant.
Grant writing is the top AI use case in the sector. TechSoup’s 2025 benchmark found that 60% of nonprofit professionals name it the use case they most want AI help with, and nearly 25% of surveyed nonprofits are already using AI to streamline grant applications and donor outreach.
AI handles the structural and mechanical parts well: reformatting narrative sections to match different word limits, rewriting boilerplate org descriptions, generating a first-draft logic model, and summarising supporting evidence. Time-savings estimates from purpose-built grant writing platforms suggest proposals can come down from 30 to 50 hours to 8 to 12 hours with AI-assisted drafting. For a small org using general AI tools for the first time, a more honest expectation is 2 to 4 hours saved per proposal in the early months. Human review remains essential for strategy, voice, and accuracy. The AI gives you a place to start. You still make the case.
Tool and cost: ChatGPT Plus at $20 per month covers one staff member doing drafting and content work.
Summarising meetings and calls
AI transcription tools turn a recorded board meeting or staff call into a draft summary in minutes, removing one of the most time-consuming administrative tasks a small-org leader handles.
Tools such as Otter.ai record and summarise calls automatically. One nonprofit leader, cited by Bridgespan, described AI meeting summaries as reliable “first drafts” that removed hours of manual note compilation. The output still needs a check, but checking a draft takes fifteen minutes rather than an hour of reconstruction from handwritten notes.
Otter.ai is available to nonprofits at 40% off via TechSoup, bringing the Business plan to roughly $18 per user per month. Google Gemini is included free in Google Workspace for Nonprofits and can summarise meeting transcripts from Google Meet recordings. An org already on Google Workspace for Nonprofits can start this workflow at near-zero cost, which makes it the most obvious first use case for many small teams.
Writing social media posts and email appeals
AI can produce a usable draft of a social post, donor email, or event announcement in under a minute, which matters most when communications staff are stretched across multiple roles.
For an org where the ED or one staff member handles all external communications, the bottleneck is usually time to draft, not ideas. AI removes that bottleneck for routine content: event announcements, end-of-year appeal drafts, newsletter intros, and social posts for upcoming campaigns. The Virtuous 2026 Nonprofit AI Adoption Report found that faster drafts and improved content quality are among the most commonly reported gains across the sector.
The realistic expectation: AI drafts are generic by default and will need editing for voice and specificity. The time saved is in getting from blank page to a starting point, not in cutting the review step. Canva Magic Studio is free for verified nonprofits and handles simple graphics alongside copy, useful for staff with no design background. Buffer offers a 50% nonprofit discount for scheduling. No donor-specific relationship content should be fully AI-generated. This use case applies to broadcast and campaign content, not one-on-one outreach.
Repurposing content you already have
AI can turn a completed grant report, a program evaluation, or a set of interview notes into donor-facing copy, social posts, or a newsletter section without starting from scratch.
This is the use case most often left off general AI lists, and it is highly relevant to a small org. Grant reports, program evaluations, impact data, staff meeting notes, and beneficiary interview summaries can all be fed into an AI tool to produce donor updates, annual report sections, or social proof content. This is different from drafting new content: the facts and evidence are already verified, so the risk of AI generating inaccurate information is lower because the tool is summarising what you gave it rather than drawing from general knowledge.
One practical constraint: only enter information that does not include personally identifying client data. Candid’s 2026 responsible AI guidance is direct on this point. Do not enter anything you would not paste into a public website. Google Gemini in Workspace or ChatGPT Team tier both handle this type of task with appropriate data protections in place.
Is there anything AI should not touch?
A few categories consistently underperform or carry real risk: crisis communications, one-on-one major donor messages, and any task that requires clean structured data you do not yet have.
Three specific areas to avoid or approach with caution:
Crisis and sensitive beneficiary communications. AI should not draft or send messages involving trauma, safety, or individual beneficiary circumstances. The 2022 National Eating Disorders Association case, in which an AI chatbot deployed for crisis support began dispensing harmful advice without adequate supervision, remains the sector’s clearest marker of what goes wrong when AI is used in emotionally sensitive contexts. Candid’s 2026 policy guidance names crisis communications as a category where AI use is usually the wrong call.
Major donor and board relationships. One-on-one outreach from an ED to a board member or major gift prospect should remain human-authored in voice and final judgment. AI can help with a first draft, but the relationship is load-bearing and the tone is not recoverable from a generic AI output. ChatGPT will write a thank-you letter that sounds fine. It will not know that this donor’s father just died. The letter still needs you.
Predictive analytics and fundraising intelligence. These tools require a clean, structured donor database. For many small orgs, the data quality is not there yet, and AI analytics applied to messy data returns noise. Legal, HR, and compliance language also falls outside the safe zone. BDO’s Nonprofit AI Risks analysis flags eligibility decisions and individual case notes as non-negotiable categories for human judgment.
What does it cost to get started?
A small nonprofit can access meaningful AI capability for $20 to $40 a month, and several of the most useful tools are free with a nonprofit verification.
Google Workspace for Nonprofits is free for eligible 501(c)(3) organisations and includes Gemini AI tools with enterprise-grade data protections. Canva Pro is free for verified nonprofits and includes AI design and copy features. ChatGPT Plus at $20 per month covers one power user, such as the communications or development lead. Adding Otter.ai at roughly $18 per user per month via TechSoup covers meeting transcription.
An org starting with the two free tools and one paid plan is looking at $20 to $40 per month total. Adding Otter.ai brings it to $50 to $60 for one or two users. One caveat worth naming clearly: TechSoup’s 2025 benchmark found that 48% of AI-powered nonprofits report higher technology-related expenses after adopting AI. Time savings are the more reliable outcome than cost savings in the first year.
The most useful question is not “which tool should we adopt.” It is “which one task is costing us the most time right now.” Start there. Run one tool on one task for one month. The answer to the broader question gets easier from there.
Frequently asked questions
Which AI tool is the best starting point for a small nonprofit?
For most small nonprofits already on Google Workspace, Gemini is the lowest-friction start because it is included at no extra cost through Google for Nonprofits. If your team is not on Google Workspace, ChatGPT Plus at $20 per month is the most flexible general-purpose option. Start with whichever tool covers the one task costing your team the most time right now.
Is it safe to put donor information or client data into an AI tool?
Not without a clear input rule in place. The working rule is: do not enter anything you would not paste into a public website. Donor records, beneficiary names, and case notes stay out of any AI tool, regardless of which plan you are on. Anonymised drafts, your own writing for editing, and public-information research are safe to use.
Do we need a policy before we start using AI tools?
You need a short written rule before any staff member uses AI on work tasks, not a formal policy document. The rule has two parts: which tool is approved and at what tier, and what information is not allowed in a prompt. One paragraph on a shared doc is enough to start. A fuller policy is worth writing once you have a few months of actual use to reflect on.
What is the difference between using the free version of ChatGPT and a paid plan?
The free tier may use your prompts to train the AI model by default, unless you turn that setting off. The paid Team tier turns training off by default and gives you a central admin layer so you can manage staff accounts. For most nonprofit work, the free tier is fine for editing your own writing. For anything touching donor information or sensitive org data, a paid business tier is the appropriate choice.
How much time can we realistically expect to save if we start using AI?
In the first few months, a more grounded estimate is two to four hours saved per grant proposal and ten to fifteen minutes per meeting summary compared with writing from notes. Time savings compound as staff get comfortable with the tools. Cost savings are less reliable in the first year. TechSoup's 2025 benchmark found that 48 percent of AI-using nonprofits report higher technology expenses after adoption, so budget for the tool cost and count the time savings as the return.
Can AI help with program reporting and impact measurement?
AI can help with the writing side of reporting, not the data side. If you have a completed program evaluation, a set of interview notes, or an impact summary, AI can turn that material into a donor update, an annual report section, or a board narrative. It cannot pull numbers from your database, spot trends in your data, or verify whether your outcome figures are accurate. The facts have to come from you.
Will AI write a grant proposal that is ready to submit?
Not without significant editing. AI handles the structural and mechanical parts well, such as reformatting narrative sections for different word limits and generating a first-draft logic model. The strategic case, the specific evidence of your org's impact, and the relationship context that fits a particular funder still require a human. Think of the AI output as a detailed outline you edit into shape, not a finished document.