ABOUT THIS EPISODE
Most continuous improvement practitioners are using AI the same way they use Google, and getting roughly the same quality of results. David Osagie-Igiehon has taken a different approach. He builds custom agents trained on lean principles, CI methodology, and the specific knowledge domains his work demands, and the gap in output quality is significant. What he describes is less about automation and more about compressing the time it takes to do rigorous work.
In this episode you’ll learn:- How an accountant became a lead lean consultant and Master Black Belt candidate (2:41)
- What it means to use AI as a collaborative tool rather than a driving system (4:19)
- How building a custom AI agent named Dominique changed the pace of research (5:27)
- Why the quality of your prompt determines the quality of your output (7:27)
- How a properly structured prompt produced a 125-page governance document (11:23)
- Whether critical thinking can survive in a world built around AI tools (17:34)
- How to create a specialized AI agent trained on CI knowledge and methodology (20:46)
- What excites and concerns a lean and AI practitioner about where this technology is heading (29:51)
- Why the human review step is the part of the AI process you cannot skip (32:48)
Keep Learning“Critical thinking would never disappear because people are superior than machines.” – David Osagie-Igiehon
David describes reaching Master Black Belt as the foundation that lets him know whether an AI output is actually correct. If you want that same grounding before trusting AI-generated analysis of your own processes, structured Six Sigma training is where that foundation gets built.
Learn More About School of Six Sigma at https://www.gembaacademy.com/school-of-six-sigma
Stay current on new courses, podcast episodes, and continuous improvement resources. Sign up for the Gemba Academy newsletter at gembaacademy.com/news
What Do You Think?If you built a custom AI agent trained specifically on your company’s lean methodology and problem-solving approach, what would you want it to know first?
