Perspectives from practice on process, AI and decision-making.
AI vendors promise to make the slow work on definitions, dashboards and baselines unnecessary. It does not become unnecessary, it gets done later, and until then the model runs on numbers nobody has checked.
Read →Most efficiency projects begin with the wrong question: what can we automate? Automation does not fix a broken process. It runs it faster.
Read →Process efficiency projects almost always stop at the process. They redesign how work flows and ignore why people do the work the way they do. A well-designed process running on top of badly designed incentives produces worse outcomes faster.
Read →84% of failed AI projects trace back to leadership decisions made before a single model was trained. The fix is a better starting point, not a better model.
Read →Most management dashboards fail one test: name a decision made in the last twelve months because of how a number moved. The failure mode is not bad data or poor visualization. The charts work. They just produce no decisions.
Read →