The Forces That Shape Al's Uneven Progress

Automation advances unevenly across tasks and roles. Understanding the friction factors that slow adoption can help leaders navigate workforce transformation strategically.

- By Will Drover and Laura Huang

As artificial intelligence has begun to write code, pilot vehicles, and even diagnose disease, narratives of machines abruptly replacing humans have taken hold. However, the real story is slower, messier, and far more uneven. Across roles and industries, Al's emergent capabilities form what's often called a "jagged frontier," where it excels at some tasks while struggling with others. As its capabilities progress, Al is racing to fully automate some tasks while slowing to a crawl or stalling entirely for others. Leaders who understand the factors that shape this nuanced reality will be able to guide strategy and talent with clarity and steady their organizations in the process.

A 2024 analysis of Al's progress by McKinsey & Co. suggests that its unevenness shows up not just across jobs but within them, at the task level. The consultancy projects that by 2030, in its midpoint scenario, demand will fall steeply for some occupations, such as office support and customer service. This is because such roles center on tasks that Al can already handle, like entering data, processing invoices, taking orders, and responding to routine customer inquiries. However, demand will rise sharply for health professionals and other STEM-related professionals whose work centers on tasks that Al struggles to replicate, such as clinical judgment, empathetic care, and complex, nonroutine problem-solving. The dividing line isn't job title — it's task makeup. But what is it about certain tasks or roles that make them more or less amenable to automation?