Your AI Problem Might Actually Be a Process Problem

But a lot of companies are learning an expensive lesson — automating work you don’t understand just helps you do the wrong thing faster. Gartner predicted that many GenAI projects would be abandoned after proof of concept because of poor data, unclear business value, and rising costs. A 2025 MIT report reached a similar conclusion: most enterprise AI efforts are not yet producing measurable P&L impact. 

Two established practices can help: GEMBA walks and Crucial Conversations. One helps leaders see reality, and the other helps people talk honestly about it. 

Go see the work 

GEMBA is a Japanese term meaning the “actual place.” In business, it means going where the work happens and where customer value is created. The idea is closely associated with the Toyota Production System and genchi genbutsu: go and see for yourself. 

Refreshed leadership advice from CEO Stephanie Mehta

Toyota executive Fujio Cho summarized the approach well: “Go see, ask why, and show respect.” 

How does it look in practice? Suppose your monthly report says customer orders are shipping late. From your office, the explanation might seem obvious — production needs to work faster. 

Walk the GEMBA and you may find something else. Operators are waiting 30 minutes for materials. Materials are waiting for purchasing, and purchasing is reacting to changing forecasts from sales.