The AI factory
When software runs the core, humans move to the edge.
Marco Iansiti and Karim Lakhani, Competing in the Age of AI
Harvard Business Review Press, 2020
Iansiti and Lakhani argue AI transformation happens in the operating model rather than the business model. Their AI factory has four parts: a data pipeline, algorithms producing predictions, an experimentation platform, and the software infrastructure embedding the loop. Take the human out of the critical path and the classic limits on scale, scope, and learning stop binding.
Their phrase is that software makes up the core of the firm while humans move to the edge. This method takes that as true and asks the question they leave open: what is the edge actually responsible for, and how would you prove it was competent?
For each workflow, what is the human to agent ratio, and who signs the quality baseline?
What limits throughput, and what is supposed to happen when the line stops?
Where on the evolution curve does each capability sit, and does that dictate build, buy, or let go?