My next vibe coding experiment: an AI Management System Advisor.
It's a small local tool that interviews a company about its business—its size, sector, AI maturity, and risk appetite—and then designs a governance structure tailored to that context, rather than handing over a generic checklist. The advisory logic runs on OpenAI's API and covers six areas based on ISO/IEC 42001, the NIST AI RMF, and the EU AI Act:
→ Strategy & governance policy
→ Guardrails
→ Human oversight & AI output validation
→ Use-case risk assessment
→ Upskilling
→ Roles & responsibilities
What impressed me most wasn't just the result, but the process. I went from an idea to a working, branded tool in a single session with Claude Code. We iterated on the advisory logic, the UI, and even the interview flow simply by talking through the changes.
From concept to working prototype in about one hour.
