This morning I finally got access to Google's new AI coding tool, and I wanted to see if it could actually solve a real problem I've been wrestling with as an instructional designer.
The Problem I Needed to Solve
I create eLearning courses for clients using Articulate Rise 360 and Storyline. My process usually involves developing learning objectives and content, then manually storyboarding how each piece should be presented. It's time-consuming, and I kept thinking: "There has to be a way AI could help with this."
So I decided to test whether I could build a tool that takes my content and learning objectives, applies solid instructional design principles, and automatically generates a storyboard (but one that maps directly to the actual content blocks available in Rise 360).
How I Approached It
I started by feeding the AI my instructional design framework (the principles I actually use when designing courses):
- Start with proper ID foundations (ADDIE model, clear objectives, learner analysis)
- Manage cognitive load (Multimedia Principle, Coherence Principle, Spatial Contiguity)
- Default to microlearning (5-15 minute chunks that busy professionals can actually complete)
- Match format to function (explanation vs. application vs. assessment)
- Consider real-world constraints (accessibility, mobile-friendliness, etc.)
The AI built a working prototype. But here's where it got interesting.
The Pivot That Made It Actually Useful
The first version recommended generic formats like "video," "simulation," or "interactive assessment." Technically correct, but not particularly helpful when I'm sitting in Rise 360 trying to build the actual course.
So I pushed back: "The format types should relate directly to Rise 360 content blocks only."
I uploaded the Rise 360 block reference guide, and the AI rebuilt the entire tool. Now instead of vague suggestions, it recommends specific Rise 360 blocks: "Use a Tabs block for this comparison," or "This needs a Process block to show the workflow," or "Deploy a Flashcard Grid for knowledge reinforcement."
That's the difference between theory and practice. The first version was pedagogically sound. The second version was immediately actionable.
What I Learned About AI-Assisted Development
Here's what struck me about working with this tool:
1. You need to know what you want. The AI didn't magically understand my workflow. I had to clearly articulate my ID principles and the constraints of my authoring tool.
2. Iteration is everything. The first version wasn't quite right. But because I could quickly say "no, make it work this way," we got to a useful solution fast (much faster than traditional development).
3. Domain expertise still matters. The AI could write the code, but I had to know instructional design principles well enough to guide it toward something pedagogically sound.
4. Specificity beats generality. When I constrained the output to actual Rise 360 blocks, the tool went from "interesting" to "I'm going to use this on my next project."
A Quick How-To: Using AI Coding Tools Effectively
If you want to try building something similar with an AI coding assistant, here's what worked for me:
Start with your expertise, not the tech: Upload documents that capture your methodology (frameworks, principles, constraints). Give the AI your actual knowledge base.
Be ruthlessly specific about your tools: Don't let the AI recommend generic solutions. If you work in Rise 360, Storyline, or any specific platform, constrain the output to match your actual workflow.
Test and push back immediately: Don't wait to refine. If the first version misses the mark, say so right away. These tools iterate fast (use that to your advantage).
Think in terms of "actionable output": Ask yourself: "Could I take this output and immediately use it in my real work?" If not, clarify what's missing.
Remember: You're the instructional designer: The AI can handle the technical implementation, but you need to guide the pedagogical direction. Your ID expertise is what makes the tool valuable.
The Bottom Line
Did the AI write all the code? Yes. Could I have built this myself? Not quickly, and not without a lot of Stack Overflow tabs open.
But more importantly: The tool actually solves my problem. It takes my content and objectives, applies evidence-based instructional design principles, and gives me a storyboard mapped to my actual authoring tool.
That's the promise of these AI coding assistants (not replacing expertise, but amplifying it). You bring the domain knowledge and design thinking. The AI handles the technical implementation. Together, you build something useful in your early morning instead of a week procrastinating on whether you should use a flash card or a sorting deck.
Now I just need to test it on a real client project and see if it holds up under fire.
Have you experimented with AI coding tools in your work? What problems are you trying to solve? I'd love to hear what you're building.