AI Won't Replace You. But It Will Help You Leave Earlier on Fridays.

I've built prompting courses before. But when I looked around at what was out there for Learning Designers specifically — not developers, not marketers, not "knowledge workers" — I kept finding the same thing: generic advice that didn't speak to our actual work.

Nobody was talking about how to prompt for a branching scenario. Or how to get AI to write a narration script that actually sounds like a human said it. Or how to generate quiz questions that don't all have one obviously correct answer and three answers that are basically just wrong in different fonts.

So I built it myself. 😃

The goal wasn't to teach LEDs to hand their work over to AI. It was the opposite — to help them move faster and smarter while staying in the driver's seat.

AI as a thought partner, a first-draft machine, a tireless SME interviewer. Not a replacement for design thinking. A way to spend less time on the parts that drain you so you have more energy for the parts that matter.

That was the vision. Here's how it actually went.

Let me set the scene.

I had Claude. 🤖

I had a large cup of coffee. 🍯

I had absolutely no idea what I was in for. 😜

Here's what I thought would happen:

Me: Good morning! "Based on this outline, build me a beautiful interactive course on prompting for Learning Designers."

Claude: [delivers perfection in one shot]

Me: [uploads to Rise, goes home early]

Here's what actually happened:

Me: "The flip cards are stretched across the screen."

Claude: [fixes it]

Me: "Still stretched across the screen." 😖

Claude: [fixes it again, tries inline-block, grid, flexbox, table-layout...]

Me: "Still stretched across the screen." 😖

We got there. Eventually. With display: flex; flex-wrap: nowrap; flex: 1 1 0.

But here's the thing — that IS the lesson.

What All Those Revisions Actually Taught Me

Your first prompt is always a draft. I knew this. I taught it in the course. And then I lived it at 11pm staring at a copy button that kept scrolling me to the top of the page.

The fix was three words: preventScroll: true. 30 minutes of debugging. That's not failure — that's iteration doing its job.

Specific feedback moves faster than vague feedback. "Why are the flip cards now stretched across the screen (me saying you broke it) 😂" got me nowhere. 

"The flip cards are stretching to fill the full panel width instead of sizing to their content" got me somewhere. Every time I named the actual problem, the next output was closer.

Iteration isn't a sign something went wrong. It's proof something is getting better. The course we ended up with — five modules, four video wrappers, fifteen use case templates, a capstone lab, and a prompt checklist card — is genuinely was better than anything I specified upfront. I didn't know what I wanted until I saw what I didn't want. That's not a prompting problem. That's just design.

The Move Nobody Talks About

Here's the thing I wish someone had told me before I started: when you finish building something with AI, ask it how you could have prompted better.

Seriously. Just say:

"Looking back at everything we built together, what could I have prompted you more effectively to get there faster?"

What comes back isn't a list of complaints (well maybe a little) — it's a personalized coaching session based on your actual project, your actual gaps, and the specific places where things got stuck. It's like a project retrospective, except your collaborator has perfect recall of every exchange and zero feelings about being honest.

I did this after we finished the course. Claude pointed out that I rarely described what I saw — I described what I thought was wrong. Big difference. "The cards are still spanning the full width of the panel instead of sitting side by side" tells AI exactly where to look.

That one insight alone would have saved us at least four rounds on the flip cards. Maybe five. (Definitely five.)

The best part? You can do this after any project — a single prompt, a storyboard, a quiz bank. Ask for the debrief. It costs you one prompt and it makes your next project meaningfully faster.

You Don't Have to Be a Programmer. You Just Have to Ask Right.

Here's the thing nobody tells you — you don't need to know what preventScroll: true means. You just need to know how to describe your problem clearly enough for AI to figure it out.

A few prompts that saved me hours and require zero technical knowledge:

When something looks wrong visually:

"The [cards/buttons/layout] aren't displaying correctly — they're [too wide / stacked on top of each other / cut off]. Can you fix the layout so they sit evenly side by side in a single row?"

When a button or interaction isn't working in Rise:

"This is inside a Rise 360 Code Block. The [copy button / navigation / flip cards] stopped working. Can you check the code for anything that might not work inside a Rise iframe and fix it?"

When something worked before and now it doesn't:

"After your last change, [specific thing] stopped working. Can you look at the current code and figure out what broke without changing anything else?"

When you want something to look more polished:

"Can you make this look more professional? Keep the same structure but improve the spacing, fonts, and visual hierarchy so it feels finished."

And the one I wish I'd used from day one:

"Only change [the thing I asked about]. Leave everything else exactly as it is."

That last one. Every time. Trust me on that one. 😄

Lesson Learned - The Best Part!

Here's the best part — after going through this whole process I'm now building a personal library of prompts that work specifically for Rise 360 development.

Things like exactly how to describe a layout problem, how to ask for a Code Block fix without knowing the technical reason it broke, and how to request a design change without accidentally blowing up everything else that was working.

Think of it as my own LED prompt playbook — and the more I use it, the faster every future build gets. If you're building in Rise with AI, I'd highly recommend starting yours too.

One good prompt, saved and reused, is worth ten debugging sessions.

The productivity doesn't come from getting it right the first time. It comes from getting comfortable with iteration — naming the gap, adjusting the ask, going again.

And then asking AI what you could have done better. Because it will tell you. And it will be right. 😂

AI Prompts for Learning Designers