AI Problems Often Stem From Process Deficiencies
Companies may implement AI to solve issues that are actually rooted in poor business processes. Many AI projects have failed due to unclear business value and high costs.
Many businesses are finding that implementing artificial intelligence is not solving their problems, but rather accelerating the execution of flawed processes. Gartner predicted that up to 30 percent of generative AI projects will be abandoned after the proof-of-concept stage by the end of 2025, citing poor data, unclear business value, and rising costs.
A 2025 MIT report echoed this finding, noting that most enterprise AI efforts are not yet generating measurable profit-and-loss impact. This highlights the necessity of understanding the root causes of an issue before considering automation.
To address these challenges, two established practices are recommended: GEMBA walks and Crucial Conversations. GEMBA walks enable leaders to observe reality on the ground, while Crucial Conversations foster honest communication among employees.
GEMBA, a Japanese term meaning "the actual place," involves going to where work occurs and understanding value creation. This principle, closely associated with the Toyota Production System and the "genchi genbutsu" (go and see for yourself) method, encourages leaders to observe firsthand and ask "why."
For instance, if a monthly report indicates late customer order shipments, the common assumption might be that production has slowed. However, a GEMBA walk could reveal underlying issues such as material wait times, purchasing delays, or fluctuating sales forecasts that are the true bottlenecks in the process.