AI for Nonprofit Leaders: I Learned Prompt Engineering So You Don’t Have To

Building GPTs for Real Nonprofit Work

I didn’t learn prompt engineering because I was curious about AI.

I learned it because I needed it to work.

Late nights, trial and error, rewriting prompts over and over again—trying to get something useful out of a system that didn’t understand the stakes. This wasn’t theoretical. I was building tools for grant proposals, donor outreach, real conversations that actually mattered.

This is what using AI as a nonprofit leader actually looks like when it’s built around real work, not theory.

The question wasn’t how to master AI.

It was:
How do I save time without losing the quality and heart behind the work?

What I Thought Would Happen

I assumed once I figured it out, other nonprofit leaders would want to learn it too.

I imagined workshops where we’d trade prompts, refine language together, debate what works best. Leaders picking this up the same way we pick up any other tool—because that’s what we do.

We adapt.

What Actually Happened

That’s not what happened.

At Defy, when I showed other executive directors what I had built, the first response wasn’t curiosity.

It was immediate:

“Can you build this for me?”

At convenings, same thing.

Different organizations. Different rooms. Same response.

Even sitting at Harvard Business School with nonprofit leaders from all over the world, I didn’t hear:

“Teach me how to do this.”

I heard:

“Can you just make it work for me?”

Seeing the Pattern Clearly

At first, I thought it was hesitation.

Maybe even resistance.

But it wasn’t.

It was something else.

The Real Issue: Capacity, Not Capability

This isn’t about intelligence.

Nonprofit leaders are some of the most adaptable, resourceful people I know.

The issue is capacity.

Their days are already full—managing teams, navigating systems, responding to crises, carrying the emotional weight of the work. There’s very little unused space.

So when AI shows up and says:

“Learn this new way of thinking. Learn this new way of working.”

It’s not exciting.

It’s exhausting.

Reframing the Conversation

We’ve been asking the wrong question.

It’s not:
“How do we teach nonprofit leaders AI?”

It’s:
“Why are we building AI that requires them to learn it in the first place?”

Most conversations about AI for nonprofits focus on tools or tactics. But that misses how nonprofit leaders actually work day to day.

Most tools assume time.
They assume headspace.
They assume people want to become proficient in something new.

But leadership doesn’t work like that.

Building Tools That Fit

That shift changed how I started building.

Not tools that require learning.
Tools that fit.

Inside HeadspaceGenie, the goal isn’t to turn leaders into prompt engineers.

It’s to create something that works inside the way they already operate.

Grant writing. Donor communication. Thinking through decisions.

Not adding another system.

Removing friction from the ones that already exist.

This is what AI for nonprofit leaders should feel like — support that fits into your work, not something you have to learn from scratch.

Leadership Is Protecting Your Headspace

We’ve been taught that good leadership means learning everything.

But in practice, it’s the opposite.

Leadership is knowing what deserves your attention—and what doesn’t.

It’s protecting your headspace so you can focus on what actually matters.

Mission. People. Impact.

Sometimes the most important decision isn’t what to learn.

It’s what not to.

And finding support that works without asking you to become something else.

If you're exploring how to actually use AI in nonprofit work, you can see how this works in practice at HeadspaceGenie.

HeadspaceGenie was built to give you back space to think, not more to learn.