SAT Studio: Personal education app
The challenge:
My daughter needed SAT prep, but existing apps either overwhelm users with content or feel juvenile. I wanted to build a focused learning tool with bite-sized sessions and sophisticated design.
My approach:
I used Lovable to build a functional prototype in 3 hours. Applied cognitive load research to structure learning around 7-word sessions. I included audio pronunciation and a competitive dashboard to keep things light.
The result:
-
Functional app shipped in 3 hours
-
User testing with primary audience (my daughter)
-
Demonstrates rapid prototyping with maintained design standards
-
Integrated free resources like Magoosh's 350 SAT words with definitions, audio pronunciation, and example sentences
-
Live demo: grammar-guru-gems.lovable.app
Key takeaways:
AI tools accelerate implementation, but design quality still requires human judgment. Pro tip: Lovable has a UI prompt library I should have used upfront—would have solved my design consistency challenges.
VisAIbility Tool: AI Powered Brand Visibility
Client consulting product
The business challenge:
Traditional SEO is changing as users shift from Google to AI assistants.
Consulting clients had no way to track if ChatGPT, Perplexity, or Claude recommend their products. Brands were creating content without knowing if AI would surface it.
My approach:
I built a tracking tool using Claude Code that monitors brand mentions across AI platforms, compares competitor visibility, and identifies content gaps. Iterative development over several months with real client testing and regular design audits to maintain consistency.
The result:
-
Generated multiple consulting engagements from tool demos
-
Speaking opportunities: Moderated Women x AI panels at Google
-
Podcast appearances on AI + marketing trends
-
Positioned me as a thought leader in Generative Engine Optimization (GEO)
Key takeaway:
Build in public, learn in public. Sharing the tool early generated user feedback, client interest, and speaking opportunities.
Tool Selection Framework
Through building both products, I developed a framework for choosing AI development tools.
Claude Code: Patient, methodical. Best for complex logic and architecture.
Lovable: Creative, exploratory. Best for rapid UI prototyping and easy deployment.
The successful formula ended up simply being having a clear vision + strong design POV + willingness to critique AI outputs ruthlessly.