McKinsey partner flags 50% shift

Anu Madgavkar, a partner at the McKinsey Global Institute, told Fortune on May 21 that artificial intelligence and automation could transform up to half of professionals' work hours within five years. Fortune said the estimate points to a broad rewrite of tasks rather than a simple count of jobs lost. McKinsey's own research, published in late 2025, used similar language to argue that existing technologies could automate more than half of current U.S. work hours in theory, while adoption would take time and change the mix of work rather than mechanically erase roles. ### Where does the 50% figure come from? McKinsey published a report in November 2025 saying today's technologies could theoretically automate more than half of current U.S. work hours. The report, "Agents, robots, and us: Skill partnerships in the age of AI," said that finding reflected the scale of potential change in tasks and workflows, and "is not a forecast of job losses." (fortune.com) Anu Madgavkar leads McKinsey Global Institute research on generative AI, robots and the future of work, according to her McKinsey biography. That makes the Fortune interview less a standalone prediction than a public framing of research McKinsey had already released. ### Is McKinsey saying half of jobs disappear? McKinsey said no. The firm's report says automation potential should not be read as a direct layoff forecast because adoption depends on redesigning workflows, deploying systems and changing how people and machines work together. (mckinsey.com) McKinsey's media-center summary, published in early 2026, said the rise of AI is "far more complicated" than headlines imply and argued that automating tasks that make up more than half of U.S. work hours does not translate into wholesale worker replacement. (mckinsey.com) The distinction matters because the unit of analysis is hours and activities, not headcount alone. (mckinsey.com) ### Why focus on professionals rather than the whole workforce? Fortune's wording referred to professionals' work hours, which points to knowledge work where AI tools can draft, summarize, search, code, analyze and route information. McKinsey's broader work has repeatedly argued that generative AI reaches into business, legal, creative and STEM tasks, even when it does not remove the need for human judgment. (mckinsey.com) A McKinsey podcast summary published in 2026 described the same research in similar terms, saying 57% of current U.S. work hours could technically be automated with existing technologies and that roughly 30% of work hours may be repurposed by 2030. That framing suggests the near-term effect is likely to be reallocation of tasks inside jobs as much as elimination of jobs themselves. (fortune.com) ### What kinds of work are likely to expand if tasks are reworked? McKinsey's report says adoption will take time and that some roles will shrink while others grow or shift as new ones emerge. In practice, that points to more demand for work that connects tools to operating reality: redesigning processes, setting controls, managing data, training users and deciding where humans remain accountable. That is an inference from McKinsey's description of adoption as a partnership among people, agents and robots. (audible.com) McKinsey's Europe-and-beyond future-of-work research also said demand would rise in STEM-related, healthcare and other high-skill professions while declining in some office, production and customer-service occupations. That earlier work supports the view that AI changes the composition of work unevenly across occupations. (mckinsey.com) ### What should readers watch next? Fortune published the interview on May 21, and McKinsey's November 2025 report remains the primary document behind the claim. The next concrete test will be whether employers describe AI adoption in terms of job cuts, or in terms of redesigned workflows, repurposed hours and new implementation roles as 2026 hiring and restructuring plans emerge. (fortune.com) (mckinsey.com)