Companies that leverage AI effectively are not reducing their workforce entirely. Instead, they are reevaluating how tasks are executed. When I was an engineer at Stripe around 2017, I proposed a machine learning system that could reduce our support team size by half. I believed I was addressing the company’s largest expense: human resources. However, the head of support responded in a way I hadn’t expected: “Congratulations. You’ve automated the easy part.” This made me realize that the real issue was the workflow itself. Support agents were switching between 10 different tools, and institutional knowledge was trapped in isolated departments. Work was being handled manually, with no insight into patterns or obstacles. The greatest cost wasn’t the people—it was the flawed process.
