McKinsey boss Bob Sternfels breaks down how AI is changing consulting jobs: Non-client-facing roles are shrinking, and jobs that are growing are…

McKinsey is rapidly integrating AI, now employing 25,000 AI agents alongside 40,000 humans, with parity expected soon. This “25 squared” approach sees client-facing roles grow 25% while backend positions shrink, boosting output. AI saves millions of hours annually on routine tasks, allowing consultants to focus on complex challenges and unique human skills like judgment and creativity.

McKinsey is experiencing a workforce transformation driven by AI, with its top boss revealing that the consulting giant now employs 25,000 AI agents alongside 40,000 human employees—and expects those numbers to match by year’s end.Bob Sternfels, McKinsey’s global managing partner, shared these insights during a Consumer Electronics Show appearance in Las Vegas, outlining what he calls the “25 squared” approach. Client-facing roles—the consultants who work directly with clients—are growing by 25%. Meanwhile, non-client-facing positions have shrunk by roughly the same percentage, even as output from those roles increased 10%.“Our model has always been synonymous that growth only occurs with total head count growth. Now it’s actually splitting,” Sternfels explained. The firm can now expand its client-facing workforce while reducing backend roles, still achieving overall growth—something he described as a new paradigm for the company.

AI agents save McKinsey 1.5 million hours annually

The productivity gains have been substantial. McKinsey saved 1.5 million hours last year alone on search and synthesis work—tasks typically handled by junior employees. The firm’s AI agents have also generated 2.5 million charts in the past six months, freeing up consultants to tackle more complex problems.Sternfels said this shift means McKinsey consultants are “moving up the stack” rather than spending time on routine tasks that AI can handle efficiently.

What young professionals should focus on

For those entering the workforce, Sternfels identified three skills AI can’t replicate: the ability to set aspirations, human judgment, and genuine creativity. He emphasized that AI models work through inference—predicting the next likely step—but lack the capacity for truly orthogonal thinking that breaks existing patterns.The AI transformation also extends to hiring practices. Sternfels suggested companies should focus less on where candidates went to school and more on demonstrable skills, like reviewing a tech candidate’s GitHub portfolio rather than their university credentials.