AI Capstone: A Framework for Creative Intelligence.

Diagram of AI framework steps

What I Built—and Learned Along the Way

Over the last six weeks, I’ve been enrolled in an online class called AI-Aided Graphic Design, taught by creative powerhouse James Hurst. The course was designed to teach students how to leverage AI to enhance their creative and strategic skills, merging design thinking principles with advanced AI tools.

As someone who’s already been deep in the weeds exploring how AI intersects with creativity, this class helped me refine my process, sharpen my thinking, and formalize a framework I’d been slowly developing.

This wasn’t a one-off project or a single app. My capstone was the culmination of months of hands-on learning, building, testing, and refining. The output? A set of real tools and systems—but more importantly, a repeatable framework for using AI across any creative process.

“I learned by doing—through dialogue with the models themselves.”

The results are intuitively structured, emotionally intelligent, aesthetically consistent, and purpose-built for human use. For me, that’s what aesthetic intelligence means: not just how things look, but how they feel and function.

The Framework: Creative AI in Six Steps

This is the mindset I now bring to every problem I approach with AI:

  1. Identify the Problem
    If the challenge isn’t clear, start by using AI itself to uncover one worth solving.

  2. Get Smart
    Build a base of understanding through focused prompting and rapid synthesis.

  3. Iterate
    Use what you’ve learned to explore—visually, in code, audio, motion, or however you work.

  4. Analyze
    Revisit your goals. What’s resonating? Add AI feedback loops.

  5. Execute
    Sharpen your best ideas. Let AI help you refine and push toward production.

  6. Refine
    Loop back. Improve. Repeat. The system grows with you.

From Framework to Artifact

Applying this framework, I landed in the overlap between iteration and execution—and built a suite of functional AI tools, including a custom Retrieval-Augmented Generation (RAG) pipeline.

What Makes My RAG System Unique

  • ✅ No corpus size limit
  • 📂 Supports PDFs, CSVs, DOCX, HTML, and more
  • ⚙️ Custom ingesting, parsing, chunking, and embedding
  • 🎯 Prompt and instruction control at every stage
  • 🔍 Task-specific filtered retrieval
  • 🌐 Flexible UI/API—deploy anywhere
  • 🛡️ Built-in guardrails for content, compliance, and safety
  • 🔎 Full observability for debugging and performance tracking

Why It Matters

The class helped crystallize an approach I’d been developing on my own. It reinforced that AI isn’t just a tool for output, it’s also a tool for thinking—for testing assumptions, surfacing possibilities, and seeing patterns faster than I could alone.

Here’s what I’ve gained through this process:

  • A way to externalize my thinking using AI
  • A repeatable structure for creative problem-solving
  • A toolkit of prompts and workflows for both strategy and design
  • A higher level of confidence in combining design, code, and language

Final Thoughts

This capstone wasn’t about proving I could build something. It was about proving to myself that AI could become a collaborator I trust.

Through structured experimentation, I’ve built tools that reflect my own style of thinking, and that help me get to better, clearer, more human results—faster.

Want help building your own AI design framework?
Let’s connect.