From Mediocre to Mission-Ready: Elevating Your Nonprofit's AI Outputs

I think AI is the topic on everyone's mind right now, so let's not waste any time, and just, let's just jump straight in.

Remy, you've done incredible work in helping nonprofits reclaim their time using AI.

At Double The Donation, we deeply understand the power of that kind of transformation.

We saw it when nonprofits moved from clunky manual spreadsheets to automated matching gifts.

AI is that exact same leap forward, but just massively scaled up.

Still, a lot of folks are using ChatGPT, Gemini, and other LLMs as a glorified search engine.

So let's tackle that idea first.

What is the biggest mindset shift needed to treat AI less like a search engine and more like a junior intern? And briefly, how do you figure out which daily tasks to delegate it?

Really appreciate that question, and I think that, um, so much of the answer is exactly what you said.

I mean, I, I've heard someone refer to AI tools as a drunk intern, and I think that's actually a helpful way to think about this technology because AI tools can act in strange and unexpected ways.

Uh, so it's good to arrive at them expecting some level of weirdness and misalignment, but they also thrive on context and specificity, and they really want to please you as the user.

And in that way, they're a lot like junior interns.

And I think if we can remind ourselves of that fact and really bake it into how we collaborate with these tools and really think of working with these tools as collaboration, right? It is different from other technologies where we don't have the opportunity to shape most software using natural language, but large language models and generative AI tools, we have a lot more say in how they interact with us.

So often what I encourage, um, folks to do is remind themselves of that fact in a very physical way.

Maybe put a sticky note on your computer as you're starting to use large language models and other AI tools more, so that every time you collaborate with it, you remember this is a drunk intern, right? I wanna use it and interact with it in the way that I might interact with an intern, obviously without anthropomorphizing it, but setting the right expectations going into collaborating with an AI tool.

Um, I, I see a lot of people very excited about AI's possibility.

I'm very excited about AI's potential, but emphatically, AI's not a silver bullet, and one of the things that we have to understand is that it's going to fill in information gaps with assumptions, right?

Mm-hmm.

And so one of the things to your second question, um, that I often offer, uh, staff of nonprofits to think about is not only to give as much context as possible, and I can share later a prompting framework to that end, uh, but also to be really specific about when and how we go to AI tools for different parts of our work.

So one of the frameworks that I use quite often with organizations, I call the head, the hands, and the heart, and it basically gives nonprofits an opportunity to think about what elements of their work really depend on their head, meaning their expertise and their brainpower, uh, what really relies on their hands, meaning it's just moving information around from one system to another, and what really depends on their heart or their presence and their empathy and their, uh, their humanity.

And that framework tends to map pretty well against AI tools, which can augment our cognition and our head tasks, automation tools which can, uh, support hands tasks and, you know, rote manual labor, and then the heart tasks really usually stay human only.

And I find having done that with many, many dozens of nonprofit leaders over time, um, it often leads to revelations about the way we do work, but also where AI tools should fit in, right? We don't wanna just slap AI on everything that we do and hope that it works.

We really wanna be intentional about where we apply it and then get creative and expansive about what we're asking of the tool.

I really love how you mentioned how AI tools are collaborators, and that head, hand, heart framework and being more intentional in using these kinds of tools.

And I know that we- you've mentioned it earlier a little bit, you advocate for prompt engineering and custom assistance to avoid those robotic first drafts.