Firms must redesign their structure for AI gains, prioritizing augmentation. Ninety percent of workers in low- and middle-income countries work in companies of 10 or fewer.
Companies seeking to deploy artificial intelligence for frontline workers must redesign their entire organizational structure to achieve major productivity gains, according to experts at a recent Stanford University executive forum. Treating AI as a simple plug-and-play automation tool risks stalling progress and falling into a pattern of diminishing returns. Mohammad Akbarpour, a professor of economics at Stanford Graduate School of Business, compared the current AI shift to the introduction of electricity. Electric motors became available to replace steam-driven ones in the 1880s, but factories did not see major productivity gains until two decades later when their physical layouts were completely redesigned. "The real value did not come from better engines. It came from redesigning the whole system," Akbarpour said. "Technology changes faster than organizations." Susan Athey, a professor of economics of technology at Stanford, noted that the pace of this transition is throttled by organizational, regulatory, and infrastructural constraints. This slow adjustment is particularly visible in small businesses. Ninety percent of workers in low- and middle-income countries work in companies of 10 people or fewer. These owner-workers are not firing themselves, though AI will eventually shift supply and demand across all industries. Stanford Professor Erik Brynjolfsson warned leaders against focusing solely on using AI to automate human tasks. He calls this phenomenon "The Turing Trap," a reference to Alan Turing's 1950 concept of evaluating machine intelligence based on its ability to imitate humans. "I call it The Turing Trap and I want you to avoid this trap," Brynjolfsson said. He argued that mimicking human behavior leads to a concentration of wealth and political power, rather than genuine progress. Instead of pure automation, Brynjolfsson advocates for augmentation. While AI executes tasks efficiently, humans remain better at defining the right questions and evaluating the work. Effective AI for Management requires setting clear benchmarks around this human-machine collaboration rather than pushing for blanket automation. Corporate leaders at the forum shared how they are applying these principles. Hamid Moghadam, founder and executive chairman of Prologis, emphasized that leaders cannot stop technological disruption. They must adapt to it by listening closely to customers and maintaining the fortitude to look beyond next quarter's earnings. Alison Birdwell, CEO of Aramark Sports + Entertainment, noted that her company initially sought AI solutions to solve hiring challenges for seasonal, part-time positions. The effort evolved into an end-to-end solution that makes operations more efficient, with governance and support built directly into the workflow. Leaders must actively choose augmentation over pure automation to capture the true value of artificial intelligence. Effective AI for Executives & Strategy means establishing guardrails that keep people in the decision-making loop while deploying tools that handle execution. As AI capabilities grow, organizational redesign and strict governance will separate surviving companies from those that stall.
