Management in the age of AI

The management question about AI is usually framed wrong. It is asked as "how can AI help our people?" — which treats AI as a tool bolted onto an unchanged organisation. The more useful framing, from Iansiti and Lakhani, is: if we were designing this business today, knowing what AI can do, would we design the organisation this way at all?

Their central concept is the AI factory: a loop where activity generates data, data improves the algorithm, the improved algorithm improves the service, and the better service generates more activity. The competitive advantage is not the model — models are increasingly available to everyone. It is the loop, and the loop depends on organisation, not technology.

THE AI FACTORY LOOP

        ┌──────────► more usage ──────────┐
        │                                 │
   better service                   more data
        │                                 │
        └──── better algorithm ◄──────────┘

What breaks it:
  · data siloed per team        -> loop never closes
  · approval chain in the path  -> loop turns too slowly to compound
  · no feedback measured        -> nobody knows the loop is broken
  · "pilot" with no deployment  -> there is no loop at all

Note: the moat is the LOOP, not the model — everyone can rent the model.

This is why so many AI initiatives stall at the pilot stage. The model works; the organisation around it cannot close the loop — data is fragmented across teams, decisions need an approval chain, and nobody measures whether the deployed system is learning. The obstacle is rarely the AI. It is the organisation surrounding it.