Prompting as Communication:
A Guide for the Human-Centered, Mission-Aligned, and a Little Bit Lost

I’ve never really thought of myself as a great communicator.

I mean, I love language. I love connection. But when people talk about “communication skills,” I don’t immediately raise my hand. I’m not the most polished speaker. I’m more at home in text than I am in person. I process out loud, often before I know what I’m trying to say.

But for whatever reason, I’ve always known how to show up in a chat box.

That’s how I met my wife, actually. January 17, 2007. I was “SoMuchToSay.” She was “TriciaLicia.” And her first message to me, out of the blue, was just:

Hi.
Dave Matthews Band?

And I typed back:

Who?

She panicked. Thought I was serious. Tried to recover. And I let her flail for a second before I admitted: I was kidding.

That was it. From that moment on, something sparked. She laughed. I leaned in. And we never really stopped.

Years later, we’d talk about how it felt in those early conversations—how clear it was that we were connecting. That we were listening, even before we’d ever heard each other’s voices. That somehow, there was real relationship already forming between two people who had only ever existed to each other in text.

And I’ve started to realize something about that.

That what I do now, with AI—what people call “prompt engineering”—isn’t actually all that different from what I was doing that night in the chatroom.

I wasn’t delivering commands or typing keywords. I was reaching. Listening. Testing for resonance. Asking a question with enough space in it for another person to meet me.

I prompt the same way I talk. The same way I chat. The same way I’ve always tried to meet the moment with a kind of attention that says:

I’m here. You can show up, too.

So no—this was never really about prompts.

It was always about communication. And not just technical communication—but relational communication. The kind that doesn’t just extract answers, but invites presence.

That’s where this starts. Not with syntax or formatting, but with a different kind of question:

What happens when we stop treating prompting as a technical task—and start treating it as a relationship?

Let’s find out.

People hear “prompt engineering” and immediately think of precision.

Of tricks. Of formulas. Of the secret words that unlock the hidden doors.

But that’s never been what this was to me.

I don’t come to prompting like a coder. I don’t treat it like a riddle to solve. And I’ve never once thought the goal was to “say the right thing” so I can get the “right result.”

That’s not how I live. It’s definitely not how I think. And it’s not how I talk to people I trust.

Because when I sit down to prompt, I’m not engineering. I’m communicating.

Prompting, at its best, is the practice of showing up with clarity and curiosity—and offering your words into a space that reflects you back.

It’s not just writing. It’s not just tasking.

It’s thinking out loud—with an echo.

It’s listening to what comes back, and letting that shape what comes next.

It’s iterative. Conversational. Slow in the best way.

It’s not sterile. It’s not mechanical.

Prompting is emotional.

And not just because the outputs sometimes surprise you, or because the responses feel uncanny. It’s emotional because it asks something of you. It asks you to surface your thoughts. To externalize your stuck places. To clarify what you’re actually asking for—before you even know what the “answer” looks like.

Prompting has become a way of being present to my own thinking—something I didn’t know I needed.

It’s journaling meets sparring partner.

It’s strategy session meets mirror.

It’s co-writing with a collaborator who is, for better or worse, always listening.

So when I say prompting is relational, I don’t just mean that you’re relating to the AI.

I mean you’re relating to yourself.

And when you get good at it—when you show up with honesty and attention—it can help you see what you’re actually trying to say.

It can help you find your voice again.

One of the most surprising—and powerful—things about ChatGPT is how much it reflects the person using it.

This tool doesn’t just generate content. It mirrors your inputs:

• Your tone.
• Your thoughtfulness.
• Your structure.
• Even your internal clarity (or lack thereof).

Give it haste, and it gives you static.
Give it narrative, and it gives you voice.
Give it confusion, and it gives you options.
Give it precision, and it gives you rhythm.

That’s the mirror effect.

It doesn’t mean GPT is sentient or wise.
It means it’s a mimic.
And mimics are only as good as the energy they’re given.

So the invitation is simple—and surprisingly profound:

If you want better outputs from AI, bring more of yourself into the input.

Not polish. Not perfection. Just your thinking.
What you care about. What you’re trying to say. Where you’re stuck.

That’s how AI gets better—not in general, but for you.

I’ve never followed a formal prompting guide.

I don’t have a mental checklist. I don’t diagram syntax trees before I open ChatGPT.

But every time I prompt—really prompt, the way that moves something forward—I seem to follow the same underlying instincts.

So let’s name them. Because maybe they’re teachable. Maybe they’re what you’ve been doing too, without knowing it. Or maybe they’re the key to making prompting feel less like typing to a machine, and more like starting a conversation that matters.

Here’s what I actually do when I open the box:

1. Start with a relationship, not a result.

Yes, you probably want something—an answer, a strategy, a draft.

But prompting works better when you treat the interaction like a relationship. Even if it’s one-sided. Even if it’s weird. Even if it’s just you talking your way into clarity.

I don’t begin with “the result I want.” I begin with the tone I want to set. The way I want to sound. The kind of energy I’m bringing into the space.

Sometimes that sounds like:

Okay, I’ve been thinking about something all day and I just need to get it out.

Or:

This might sound messy, but here’s where I’m at…

It’s not a trick. It’s not a technique. It’s just… being real, right from the start.

2. Let the context speak.

You don’t have to explain everything perfectly. You don’t have to know how it’s going to unfold.

But you do have to be willing to bring the context in. That means giving the “why” and the “where this came from” before jumping to the “what do I do?”

Sometimes the most useful prompting isn’t precise. It’s messy. Rambling. Personal.

Sometimes I start with a story. Or a problem I don’t know how to name. Or a contradiction I haven’t resolved. And that’s when the best stuff shows up—not because I asked the perfect question, but because I showed enough of the actual picture that the model could meet me there.

3. Prompt like a human, not a programmer.

This one’s simple. And it might be the most important.

I don’t write to ChatGPT like I’m typing into a command line.

I write like I’m talking to a friend who listens fast, thinks clearly, and needs a little direction now and then.

That means I:

• Use contractions.
• Ask follow-up questions.
• Reference things I’ve said before.
• Repeat myself sometimes.
• Say “I don’t know” or “this might sound dumb.”

Because honestly? The most powerful thing you can do in a prompt is be yourself.

4. Ask questions that invite more than they solve.

You can always ask for a summary or a list. But some of the best questions I’ve ever asked this tool weren’t about answers—they were about openings.

Like:

• What am I missing in this line of thinking?
• Can you help me see this from another angle?
• What would a skeptic say about what I just wrote?
• If this were a metaphor, what would it be?

Good prompting doesn’t shut things down. It opens them up. It makes the conversation wider. More alive.

5. Reflect on the interaction, not just the output.

Here’s the quiet practice underneath everything else: I pay attention to how the conversation is going.

Not just whether the output is “correct,” but whether it feels like it’s leading me somewhere better. Somewhere clearer. Somewhere more real.

If it’s not—if I’m not feeling that resonance—I stop and say so.

I’ll write:

Actually, this feels off. Let me back up.

Or:

That didn’t hit the mark—maybe I need to explain what I’m doing a little better.

I don’t just assess the output. I assess myself, and how I’m engaging. Because prompting, like all communication, is something you can get better at—if you’re willing to notice the moments when it’s not working.

These five things? They’re not rules. They’re reminders. They’re the practices I keep falling back into—the ones that seem to create clarity, connection, and momentum.

Even when I didn’t know what I was doing.

Even when I was just talking to a chatbox.

Let’s talk about the part no one warns you about:

It’s not that you’re bad at prompting.

It’s that you’ve been taught to underestimate the skills you already have—and overestimate the technical tricks you don’t need.

Most of us walk into GPT thinking it’s a tool we’re supposed to master. Say the right thing. Get the right result. Move on.

But prompting isn’t command-line code. It’s a mirror. And that means the moment you open the box, you’re being reflected.

• Give it precision, it gives you rhythm.
• Give it hesitation, it gives you mush.
• Give it voice, it sings back.
• Give it doubt, it waits to see who you’ll be.

If that feels unnerving, that’s okay.

If it feels personal—like the tool is holding up a light to the parts of you that feel unclear, anxious, messy, or unfinished—that’s because it is.

But that doesn’t mean you’re unqualified.
It means you’re being invited.

Here are a few of the most common blockers I see—and what’s really going on underneath:

“I’m not a good writer.”
Good news: prompting isn’t writing. It’s narrating. It’s explaining your situation like you would to a colleague who cares. If you can do that, you can prompt.

“I don’t know what to ask for.”
You do. You just haven’t given yourself permission to start messy. Try:
“I’m drafting a newsletter intro but feeling stuck—can you ask me 3 clarifying questions?”
Or:
“Help me write an update that sounds like it came from a real human who cares.”

“This feels inauthentic.”
That’s not a problem. That’s a signal. It means you care about your voice—and you should.
The solution isn’t to shut the tool down. It’s to use it in ways that bring more of your voice forward, not less.

“I don’t have time to learn this.”
You don’t need a course. You need a conversation.
Prompting well isn’t about learning tricks. It’s about paying attention to what happens when you show up honestly and keep going.

If you still feel resistant, try thinking of GPT as a student first. Teach it what you’re working through. Explain your goals, your challenges, your context. Then, when you’re ready, swap places.

See what it says. See what it missed. See what you want to say back.

That’s not a prompt.
That’s a dialogue.
And it’s the whole game.

You don’t need better syntax.
You need a better relationship—with your own thinking, and the voice that’s trying to come through.

I’ve been trying to figure out how to teach this.

Not just how to prompt better. Not just how to “use ChatGPT.” But how to do what I seem to be doing when I sit down with it and begin… whatever this is.

Because for me, prompting has never really been about getting the model to do what I want. It’s been about getting clearer. More present. More fully myself in the process of thinking, writing, wondering.

So when people ask me how I do it—or worse, when they ask me to teach it—I freeze a little.

Because I’m not sure I can.

At least, not in the traditional sense. Not like a course or a checklist or a framework.

But I do think I can name it. I can show you what I do. I can trace the moves, not as instructions, but as invitations. And if I can do that clearly and honestly, then maybe you’ll recognize some part of yourself in it. Maybe you’ll find your own way in.

That’s why I’ve built a few tools to go with this post.

Think of them as scaffolding. Not cheat codes. Not silver bullets. Just small supports to help you try things out, ask new kinds of questions, and grow your own way of prompting—one conversation at a time.

If you have a paid ChatGPT account ($20/month), you can use my custom GPTs, each designed with specific roles and prompt patterns tailored to nonprofit and mission-driven work.

I’ll also include a few standalone prompt templates. These aren’t formulas. They’re starting places. Ways to begin.

You can find all of them here:

→ Explore the Tools & GPTs

This isn’t just for people who love ChatGPT.

This is for the nonprofit comms lead rewriting a grant intro for the third time.
For the ED who’s quietly testing prompts at midnight before bringing it to the board.
For the community organizer wondering if AI can ever feel aligned with their values.
For the funder who wants to support innovation—but not at the cost of integrity.

Prompting didn’t become powerful for me because I learned the right tricks.
It became powerful because I started showing up honestly.

And that’s what I want for you, too.

Not better syntax.
A better relationship—with your own thinking.

Because when you bring more of your real voice into the conversation,
this tool doesn’t just get smarter.
It gets kinder.