McKinsey says it has 25,000 AI agents. Its rivals say that's not a metric of success.

In a bold statement last month, McKinsey CEO Bob Sternfels revealed that the consulting firm has integrated 25,000 AI agents into its operations in under two years. However, this figure has sparked skepticism among competitors. Steve Newman, EY's global engineering chief, expressed doubts about the significance of such a number, stating, "I don't think the number of agents translates directly to value." He emphasized that the real impact often comes from a select few agents that contribute significantly to overall productivity. Newman highlighted EY's focus on efficiency, explaining that the firm evaluates agent performance through key metrics related to productivity, quality, and cost. "We chart those month to month, quarter to quarter, and that's what my lens is as to the effectiveness of AI," he added. Similarly, PwC's chief AI officer, Dan Priest, shared his perspective, asserting that the emphasis should be on the quality of AI agents rather than their quantity. "I think that's probably the wrong measure," he remarked, reinforcing the notion that successful AI deployment is defined by effectiveness, not mere numbers. Sternfels originally announced McKinsey's AI agent onboarding at the Consumer Electronics Show in Las Vegas earlier this year and reiterated this during a Harvard Business Review podcast. A McKinsey spokesperson later confirmed the accuracy of the 25,000 figure, with Sternfels indicating plans to expand this number further. He noted that within the next 18 months, every one of the firm's 40,000 human employees would be supported by at least one AI agent. The rise of AI technology has significantly transformed the consulting landscape, prompting firms like McKinsey, EY, and PwC to rapidly adapt and position themselves as leaders in AI integration. EY revealed during an October earnings call that it invests over $1 billion annually to create AI-centric platforms, which includes the development of 1,000 AI agents and the deployment of more than 100 internal AI applications.