06: Stop Telling AI What to Do. Start Showing It How to Feel.

(5 Designer Prompting Principles)

📺 Prefer video?

Watch the 6.5-minute version, or keep reading

(This is a good one to share—forward it to someone who’s been wrestling with robotic AI outputs.)

---

Where We Are

A head of school uses ChatGPT to compose the weekly memo for parents.

The result is grammatically perfect but entirely forgettable.

She sighs, concluding that AI can’t capture her school’s unique voice.

But it can.

The problem isn’t the AI; it’s how we prompt it. We often use AI like dull scissors, hoping for precision but settling for mangled output.

---

Why Most AI Prompts Produce Robotic Results

Most of us use AI like a vending machine: “Write an email about spirit week.” “Create a parent survey.” “Draft a staff memo.”

Press button, get output.

And that’s exactly what we get: functional, forgettable, and robotic results.

But if you treat AI less like a task-taker and more like a communications partner, it starts producing work that actually sounds like you. No complex prompts required. Just a shift in how you frame what you need.

The shift? Think like a designer instead of a task manager.

---

What Designer Prompting Actually Delivers

Designer prompting applies the same human-centered thinking you’d use when briefing a new hire, focusing on people, feelings, and context rather than just deliverables.

Here’s what changes:

Fewer revisions: Better prompts mean less back-and-forth editing.

Consistent voice: AI that understands your tone produces communications that sound authentically yours.

Faster creation: When AI grasps your context, content happens with less effort.

---

The Five Principles of Designer Prompting

So what does designer prompting actually look like?

Instead of listing requirements, you describe the experience you want to create.

Instead of specifying outputs, you paint the scene:

• Who’s reading this?

• What are they feeling?

• What do they need?

Here are five principles that make this shift concrete, turning vague AI into a communications partner that gets your school.

Principle 1: Start with Feeling, Not Features

Before you mention format or length, define the emotional energy.

Instead of: “Write a message about the new homework policy.”

Try: “Write a message about our new homework policy. It should feel like a calm, confident conversation—not defensive, not apologetic. Warm but authoritative.”

AI tools are surprisingly good at capturing tone when you give them an emotional target.

Principle 2: Design the Scene, Not Just the Output

Describe what happens when someone encounters your communication.

Instead of: “Create a staff memo about the crisis protocol.”

Try: “When teachers read this, they should immediately know: what changed, why it matters for their classroom, and what to do in the next 24 hours. After reading, they should feel more prepared, not more anxious.”

You’re designing a reading experience, not just delivering information.

Principle 3: Prompt for Behavior

Design for completion behavior and emotional response.

For a parent survey: “The survey should feel quick: parents should think ‘this will take 2 minutes’ when they see it. The tone should communicate ‘we genuinely want to improve’ not ‘we’re checking a box.’”

Principle 4: Acknowledge the Emotional Moment

Map the emotional landscape first.

For standardized testing week: “Parents receiving this are probably feeling stress about their child’s performance, confusion about what scores mean, and worry about test anxiety. The email should acknowledge these feelings without dwelling on them.”

When you name the emotion, AI can write to it rather than around it.

Principle 5: Build Context Once, Use It Forever

AI has conversational memory. Use it.

Start your conversation with: “I’m the communications director at an international K-12 school with 650 students. Our communication style is warm but clear; we avoid jargon and write in short, scannable sections.”

Now every subsequent prompt benefits from this foundation.

---

What This Looks Like in Practice

Let me show you the difference.

Task-Manager Prompt: “Write an announcement about spirit week.”

Result? Generic. Forgettable. Doesn’t inspire participation.

Designer Prompt: “Write a spirit week announcement for middle school families. It should feel like an invitation to fun, not a formal notification. Open with the most exciting theme to hook attention. Use short, energetic sentences. Close with something that makes participation feel easy and inclusive.”

Result? Completely different energy. Achieves the goal: high participation.

Same AI. Different prompt. Different result.

---

Your Move This Week

Pick one communication you send regularly.

Rewrite your prompt using these principles:

1. Start with feeling: “This should feel like...”

2. Design the scene: “When someone reads this, they should...”

3. Prompt for behavior: “After reading, they’ll...”

4. Acknowledge emotion: “The reader is probably feeling...”

5. Build context: “We’re a [type] school with...”

Compare the outputs. Forward both to one colleague and ask: “Which one sounds more like us?”

The difference between “make this better” and prompting like a designer is the difference between spending 2 hours rewriting and spending 10 minutes refining.

That time compounds weekly.

---

Your Experience

Have you tried prompting AI for school communications? What’s worked? What’s made you want to throw your laptop out the window?

Hit reply. I’d love to learn more with you.

And if you haven’t filled out that 10-second survey I sent last week, it’s still open—your input shapes what I build next. 🙏🏻

Systematically yours,

---

About the Author:

G (short for Gitane) is co-founder and Chief Creative Officer at EKG Collective, helping international schools turn communication complexity into systematic clarity. Learn more at ekgcollective.com.