Q&A: How Small Nonprofits Can Use AI on Purpose, Without a Tech Team or Extra Budget

1. What is one practical thing a nonprofit with three or fewer staff can try with AI this month using only free tools, and how would they do it?

The number one tip I would give a small, lean nonprofit on how to use AI this month is simply to start. Many people avoid using AI because it’s unknown: you aren’t sure how to use it, you aren’t sure what to use it for, and you aren’t sure if you really trust it.

The minimum: use it like you would have used Google five years ago.

This is not revolutionary, but it’s practical for increasing adoption. If you had super powers but never used them (or didn’t know how), they wouldn’t be helpful. You would have to learn how to use them through trial and error and repetition. In today’s world, you literally have super powers (more specifically, super intelligence) in the palm of your hand, but if you don’t use it, you’ll never gain the benefits.

From there, get in the habit of asking more of it (literally). Start to use it for conversation, thought companionship, and thinking deeply through any practical challenges you are facing in your day to day.

Tactical example: If you want one specific thing to try this month, pick something you already have to write, like a donor thank-you letter, a volunteer recruitment post, or even a grant. Open the free version of ChatGPT or Claude. Tell it who you are, who the message is for, and what you want it to do. Paste in an old example of one you liked so it can match your voice (leave out anyone’s private information). Ask for three versions. Then edit the best one until it sounds like you. That takes 15 minutes, and you will learn more from it than from reading about AI for a week.

To summarize, just start!

Hack: Ask AI how to use AI! If you like videos, ask for videos. If you like text, it can provide that. You can even put it in voice mode if you want to chat about it. Ask it for examples, images, or whatever you can dream up. It will most likely be able to do it.

2. What is the biggest misunderstanding you see nonprofit leaders have about how these tools work?

The biggest thing I recommend to nonprofits, and anyone just getting started in learning how to use AI, is to spend some time understanding what it is and what it isn’t. Once you understand a little more about how the technology works, it removes the fear of using it. Using AI is a great way to do this as well.

Many are afraid to start because there is a natural human instinct to fear the unknown. It’s a trait that has kept humanity safe for thousands of years. But once you realize this little box that gives us back words is simply another piece of technology, it disarms the fear.

The misunderstanding I see most is treating it like either magic or a search engine that looks up facts. It is neither. It is predicting the most likely next words based on everything it has learned. That is why it is so good at writing and thinking with you, and also why it can say something wrong with total confidence. Use it, and check what matters.

3. Which free AI tools are reliable enough for a small nonprofit to depend on, and when is it worth paying for something instead?

I highly recommend optimizing toward simplicity to help increase your adoption. For simplicity, I would orient toward one of the primary frontier models: OpenAI’s ChatGPT or Anthropic’s Claude.

These two organizations and platforms are leading in the AI model space. They are consistent and reliable. If your organization uses Google Workspace, Gemini may be a great option, and if your organization uses Microsoft, Microsoft Copilot may be great to use as well.

The primary practical difference is that Gemini and Copilot are commonly used inside the platforms you use day to day in Google and Microsoft, respectively. This is a great way to use AI, but if you are looking to use AI widely for all activities, I would highly recommend ChatGPT or Claude.

Even with tight budgets, if you can justify the $20/account subscription for ChatGPT or Claude, I believe that is a great bang for your buck. But before you pay full price, check the nonprofit programs listed at the end of this article. Several of these companies offer free or heavily discounted plans for nonprofits.

An important thing to note: under most platforms’ terms of service, they are allowed to train on what you put into free plans and individually paid plans. You do have the option, on any plan (free and up), to turn off training in the settings. If you have a business or team subscription, they do not train on your data by default, per their terms.

So where is the line on paying? Free is plenty to get started. I would pay once you are using it every day and running into the limits of the free version, or once your team wants to work on it together. What data belongs in these tools and what doesn’t is a different line, and that one is yours to draw as a leader in your organization. You know your donors, your clients, and your risk tolerance better than I do.

The other free tool I would highly recommend is not actually an AI tool: YouTube, or what I like to call YouTube University. YouTube has incredibly simple and practical videos that can walk you through the basics of how to use AI.

4. If a nonprofit with no legal team and no technical staff wants to write a basic AI use policy, what should the first line say?

A simple but practical way to start would be something like this:

“We encourage our team to use AI to do better work for the people we serve. Whoever uses it is responsible for what it produces, and we do not put private information about donors, clients, or staff into any AI tool the organization has not approved.”

Make it your own and put in any specifics that fit, but this is a simple way to enable your team to begin benefiting from the technology while keeping sensitive information out.

More often than not, policy tends to restrict rather than promote. I would challenge you to look at an AI policy as a way to give permission to the greater organization to use AI in a way that you know the organization is aligned around, will be effective, and matches the governance for your organization.

One thing that will be true is that this policy will change, maybe even just three months from now, and that is okay. Traditionally, nonprofits and businesses operated in a relatively stable environment. Technology and markets took significant time to change. Today, the pace of technology development and the market’s response to it can change by the hour, so we have to be able to adapt and be agile in an ever-changing world. Expect this, be agile, and consistently adapt.

5. Who in a small organization with no dedicated IT or compliance role should own that policy once it exists, and what do they need to do day to day to keep the organization safe?

I wouldn’t recommend putting this on one person’s shoulders. AI is technology that needs to be understood and adopted organization-wide. Yes, one individual should be responsible for executing the policy, but for the policy to be effective, especially in a small organization, it should feel like a shared agreement across the staff.

As for who in a small organization should own it, I would recommend a step-by-step process like this:

1. Form a small committee to represent the organization. Call it the “AI committee.” It should be made up of strategic members who (1) understand the organization’s overall goals, initiatives, and risk tolerance, and (2) understand AI well enough to make sure the organization gets significant value out of the technology while identifying opportunities to apply it throughout the organization.
2. Within that committee, assign one individual to be responsible for drafting the first version of the policy, with the overall organization’s strategy and risk tolerance in mind.
3. That individual should spend time talking with the leadership team, the committee, and the greater organization to understand how the team uses AI today and how they would ultimately like to use it. They should also share how they are thinking about the policy and get feedback.
4. From there, the policy writer should draft the AI policy so that it aligns with the goals of the organization, gives practical ways for the team to use AI effectively within the organization’s risk tolerance, and reflects the AI committee’s view of the greater good of the organization.

If you are a team of three, the committee is the whole staff. Put 60 minutes on the calendar, have the conversation together, and develop your point of view as a team.

Day to day, the individual responsible for the policy should understand the tools and technology the organization is using today, have direction on what tools the organization should be using, and understand what guardrails, if any, should apply to the organization’s AI usage. Things to consider: What tier of service are you on? Can companies train on your data? What data are you comfortable putting into AI?

6. Much of the software nonprofits already use, including email platforms, CRMs, and accounting tools, now advertises “AI-powered” features. What should a non-technical person ask before trusting one of them with real data?

You do not need to be technical to ask good questions. Email your vendor’s support team these five and see what comes back:

1. Is this AI feature on by default, and can we turn it off?
2. Is our data used to train your AI, or anyone else’s?
3. Which AI company is behind this feature, and does our data leave your system to get to them?
4. Who can see what the AI produces, and does it respect the permissions we already set?
5. When the AI gets something wrong, how would we know, and how do we fix it?

A good vendor answers these quickly and in plain English. If they can’t or won’t, that is your answer. Leave the feature off until they can.

Another hack is to ask AI these questions and reference the organization’s terms, privacy policies, and support documentation. That way, you can be informed going into a dialogue with the vendor.

7. What mistakes do you see organizations make when they start using AI tools?

There are two types of organizations I see: organizations with low adoption and organizations with high adoption. Each has its own mistakes.

Lower-adoption organizations

Quitting after one try. The mistake is trying it once, getting an average answer, and deciding it doesn’t work. Using AI is a skill. You get better at it by using it, the same way you got better at searching Google.

Making it one person’s job. The mistake is handing AI to one person, usually whoever is “good with computers,” and assuming the organization is covered. One person using AI is not company adoption. It only pays off when the whole team is using it in their own work.

Higher-adoption organizations

When an organization has high adoption rates, they tend to try to “AI” everything. Remember, AI is a tool to do your work more effectively. It is a means to an end. AI in itself is not the point. Make sure you’re applying AI where appropriate, and not spending more time using AI just to use AI and actually becoming less effective.

All organizations

Often, with this newly unlocked ability to produce an incredibly high output of high-quality material, early adopters forget to apply the human judgment step.

Human judgment and taste are the strongest tools when it comes to using AI effectively.

8. For a nonprofit with no training budget, what is the most realistic way for staff to build their own AI skills on their own time, without paying for a course or program?

As mentioned before, my first recommendation is to use AI to learn AI. From there, YouTube is an incredible free resource to find anything you want, from 101 to PhD-level education on AI. Any major model provider will have significant education, resources, and up-to-date blogs and releases you can continue to learn from: OpenAI, Anthropic, Google, SpaceXAI, etc.

I would recommend 15 to 20 minutes a day of self-education. Then have your AI committee put on a simple education session or roundtable every month or so, at whatever cadence fits your organization, where everyone shares what they are learning. This shared, applied learning is a different type of learning than self-education. It is effective for learning, and it also sparks ideas that could be valuable for your organization.

If you want something more structured, here are two free places to start:

• AI Fluency for Nonprofits, a free course from Anthropic and GivingTuesday built for people who are new to AI with no technical background.
• OpenAI Academy, which has free workshops and tutorials.

Also look at local events. For example, AI Indy hosts many free events, from workshops to meetups, education sessions, and panels. What I have found is that being in person allows someone to gain practical skills and experience with technology at a much faster pace than online, self-guided tutorials.

Nonprofit AI programs by provider

Programs and prices as listed by each company in early October 2026. They change, so confirm on the company’s site.

About Sterling Curran and AI Indy

Sterling Curran, Founder + CEO of AI Indy, Founder + CEO of Heartland Data Co, and Co-Founder of SalesAi

Sterling Curran has spent the last six years building and running technology companies in Indianapolis. He started in a commission-only sales seat with a biology and exercise physiology degree and no prior business, technology, or sales experience. From there he worked his way up to running overall company operations. He then co-founded SalesAi, a conversational AI company he led and grew rapidly from $0 to $6M ARR in 8 months while becoming a market leader in the early voice technology market, with AI agents running inside hundreds of customer businesses. Putting AI to work in that many organizations taught him that the technology is moving faster than most people and teams can keep up with. He founded AI Indy so the people in his community are not left behind by tools that are developing fast. He started Heartland Data Co in the same spirit as AI Indy, to help the Heartland and Midwest capture their share of the value the AI era is creating. Outside of work, he trains for and competes in Ironman triathlons, marathons, and HYROX, and loves anything physically active.

AI Indy: Sterling started AI Indy after traveling the country for SalesAi and finding communities everywhere having excited, productive conversations about AI, then coming home to Indiana and not finding the same energy and abundance. He founded it in 2025 to make sure Hoosiers don’t get left behind in the AI era. AI Indy helps increase AI adoption, competency, and education, especially for people who feel like they are behind. It does that through meetups, panels, and hands-on workshops built around practical use. Nonprofit staff and leaders can engage by coming to an event, joining a workshop to build skills alongside peers, or signing up for updates at aiindy.com.

To learn more about AI Indy, sign up for the newsletter at aiindy.com, which shares upcoming (typically free) events to attend. To engage directly, email Sterling at Sterling@Sterlingcurran.com.

About Jeff Bennett

Jeff Bennett, chief innovation officer, brings nearly 30 years of civic, nonprofit, and government leadership experience to CICF. He develops new partnerships and advances CICF’s regional impact. Prior to joining CICF, Jeff served in key leadership roles for the City of Indianapolis, where his work in neighborhood development, housing stability, human services, and pandemic response helped shape long-term strategies to support residents and strengthen communities.

“SI” is the new federal term for “AI.” On September 29, the President signed an executive order directing federal agencies to use “Super Intelligence,” or “SI,” in place of “AI.” If you see SI in a grant notice or a federal document, it is referring to the tools you already know. For now, the federal definition is the same, and a formal definition is still being developed.

The big labs made safety commitments. The same day, six leading AI companies signed a voluntary White House accord with four layers: internal safety controls, internal teams to oversee them, outside auditors, and board-level oversight. It is voluntary, not law, but it is a public commitment.

Know the names. OpenAI makes ChatGPT. Anthropic makes Claude. Google makes Gemini. Microsoft makes Copilot. Meta makes Meta AI. SpaceXAI (formerly xAI) makes Grok.

How it works. These tools learned patterns from an enormous amount of writing, and they answer by predicting what words should come next. That makes them great at drafting, summarizing, and thinking out loud with you. It also means they can sound confident and be wrong, so check anything that matters.

Nonprofits get a deal. Most of these companies offer free or discounted plans for nonprofits. Details are below.

You stay responsible. AI helps you do the work. It does not own the result. You do.

We’d love to hear your questions, ideas, and feedback. Use the form below to send a message directly to this article’s author.