Commentary: AI agents need middle managers too

Middle managers may soon find their roles transformed as artificial intelligence agents handle routine tasks, according to Kai-Fu Lee, CEO of 01.AI and former president of Google China. In a piece for the Financial Times, Lee argues the necessity of middle management arose from the need to process and make sense of vast amounts of corporate information.

However, with AI capable of accessing and interpreting "dormant data" such as meeting transcripts, support tickets and shared documents, the need for middle managers could diminish.

Lee envisions a future where organisations adopt a shorter chain of command, with executives at the top and directly responsible individuals (DRIs) further down. DRIs would own outcomes from start to finish, such as products, markets or major risks. Middle managers would shift from overseeing execution to building trust, fostering relationships and guiding teams through uncertainty. This change would amplify the value of human qualities, such as judgement, mentoring, and navigating complex relationships.

While AI agents can execute routine tasks faster and more cheaply, Lee warns that flattening organisations too quickly may lead to frayed coordination, unhappy staff, and burnt-out executives. Moreover, experienced managers possess institutional knowledge gained from past failures and relationship challenges, a wisdom that is irreplaceable by AI. As AI takes on more tasks, the role of middle managers will evolve, with their primary focus shifting towards human-centric skills.