Part IV — Decide What the Organization Does with the Capacity AI Releases
The visible result of AI is usually speed. A draft appears sooner. A search takes less effort. A customer-service representative begins a reply with context already assembled. A manager who once spent half a day consolidating updates can see a first version in minutes.
That visibility invites a premature conclusion: if the work is faster, the organization needs fewer people. But speed is not yet capacity, and capacity is not yet a staffing decision. Between those statements lies the work that determines whether an organization becomes stronger or merely thinner.
AI changes the economics of execution. It can reduce the effort required to produce a first answer, complete a routine check, organize information, or prepare a standard response. It does not, by itself, decide what the freed effort should buy. It does not preserve the judgment that made the old work safe. It does not tell employees why they should contribute the exceptions, workarounds, and customer knowledge that make a workflow real. And it cannot choose which trade-offs the organization is prepared to live with.
Those are management decisions. This part takes them in order. First, do not confuse a released task with a layoff plan. Then give the released capacity a destination and a test. Build the trust that makes people willing to supply the judgment the system cannot invent. Finally, make clear which decisions belong to the CEO, even when the work of carrying them out belongs elsewhere.