AI reduces the cost of accessing many specialized capabilities. A generalist can research a market, model data, explore design directions, prototype software, draft across formats, and operate workflows at a level that previously required several handoffs. This does not eliminate expertise. It changes how expertise is assembled.
Research on generative AI at work supports part of the leverage thesis while also showing its limits. In customer support, AI assistance produced the largest productivity and quality gains for less-experienced workers by helping diffuse useful patterns. Experienced workers saw smaller gains and some quality tradeoffs. The conclusion should not be “AI makes everyone equally expert.” It should be that capability can move across a workforce in new ways, and system design matters.
