As generative AI makes its way into more and more of the tools we use every day, nonprofit professionals may be asking themselves whether and when it makes sense to use AI-powered features. For grantseekers and fundraisers, the answer may be nuanced. AI may make sense in parts of your workflow, but as always, a dose of care is in order.
Phil Ferrante-Roseberry is a member of the development team here at TechSoup. A self-described AI "tinkerer," Phil has found ways to integrate AI tools into his workflow over the past year. But he picks and chooses his spots for when and how to use them.
"I am pretty selective and judicious about my use of AI in development, in large part because I know that every time I touch it, it's got a climate impact," Phil says. "It's kind of the same approach that I take to getting on airplanes. I do it, but it better be worth the impact of me doing that. It's not something I use frivolously."
Business-Grade AI Is a Must
Phil primarily uses Microsoft 365 Copilot in his work, a private, business-grade instance of Microsoft's AI-powered featureset. He uses it in lieu of the free chatbot because of the way it integrates with other business tools, such as email.
"One of the reasons I'm primarily using Copilot is because my configuration is such that Copilot, every time I touch it, has access to pretty much everything on the network that I have access to," Phil says. "So all of my e-mail going back decades now, it can pull from everything on SharePoint I can access, including pretty much every proposal that we've ever issued out into the world. So it starts with a whole base of information."
This surfaces information that he may not have thought about or remembered so he can use that context to inform new grantseeking opportunities and proposals.
In addition, business-grade AI tools generally come with better data privacy: With Microsoft 365 Copilot, your organization's internal data stays private to your organization, and isn't used to help train models. This is especially important if you're working with personally identifiable information, which you should never enter into a free, consumer-grade AI chatbot.
A Potentially Useful Research and Brainstorming Tool
Phil finds this ability to pull from and reference internal information especially helpful in the context of researching potential funders and developing concept notes for grants.
"It's going off and pulling things from my own past that I either am very aware of or that I'd forgotten, Phil says. "[A]nd it'll go off and reference concept notes that we've written in the past and it's like, OK, so this is like this other thing that you wrote six months ago."
With that in mind, Phil may ask Copilot's built-in Researcher agent to do a little cursory background research to determine whether a potential funder is a good fit for the work we do at TechSoup. Copilot will then put together a summary of that funder and its activities based on publicly accessible information online — as well as internal documents and emails.
Example Prompt
Here's an example of a prompt that you can use as the basis for your own research into funders:
Tell me about this funder's activities over the last 3-5 years. How do they fit with our mission and goals?
He notes that it sometimes takes a little nursing along to get Copilot to provide all the information that he's looking for.
"I'll take that first response, and it's usually pretty good. And then I'll try to dial it in a little bit more. 'Oh, tell me about this [initiative] that they've had going[.] Who are their big grantees? What is their granting history with them?' And so on and so on."
He may then also use it to develop a first draft of a concept note, using a grantwriter agent, which he'll then use as the basis for his final draft.
Example Prompt
Here's a prompt for a fictitious nonprofit's concept note that you can adapt to your needs:
Help me write a concept note for a grant proposal. We're a small nonprofit in the Portland Metro area that brings therapy dogs to those in long-term care facilities. We are seeking a grant of $5,000 that will pay for new supplies for our animals.
You can improve upon this by providing more details about your organization — your mission, your programs, more details of how you'll use the money, and so on. The AI tool you use may also ask you for additional information. Depending on the tool you're using, you may need to be careful about divulging sensitive or confidential information.
A Helping Hand When Used Thoughtfully
As we've always said at TechSoup, generative AI can be a helpful tool when used responsibly. So with that in mind, here are a few additional tips.
- Make sure you have an AI acceptable use policy in place.
- Be transparent about your use of generative AI in producing content.
- Use a business-grade AI tool if you can. When choosing a tool, be sure that it keeps your organization's data private and that it won't use it for training purposes. And make sure that it meets your organization's requirements for data privacy. TechSoup offers Microsoft 365 and the Copilot add-on to eligible nonprofits.
- If you can't use a business-grade AI tool, be judicious about what you enter into a free chatbot like the free versions of ChatGPT, Copilot, or Google Gemini. Do not enter any sensitive information into these tools, including any personally identifiable or confidential information. Unless the terms of use or service contract stipulate otherwise, many of these free tools will use your data to train the chatbot. Again, you may need to consider your organization's data privacy requirements here.
- Be sure to fact-check any information that a chatbot provides. Chatbots are known to occasionally fabricate information, a phenomenon known as hallucination.
- Don't use an AI-generated draft as your final draft. At the very least, review and revise it first so it fits our organization's voice and that it meets your needs.
And finally, don't be afraid to experiment and try new things with AI. Good luck!
[Thumbnail photo: Shutterstock]
