The Org Chart Makes its Comeback
From hierarchical control to dynamic possibilities and back – with AI at the helm, what was out is now in again…
I was so wrong.
Ever since Jane and my time at Mozilla, the global non-profit open-source organization that brought the Firefox web browser to the world, I have been talking about “chaordic systems.” Chaordic systems, or “chaord” for short, were invented by Dee Hock, the genius executive behind the VISA global payment system, and describe a self-organizing structure that blends chaos and order to enable decentralized adaptability, resilience, and emergent innovation. As this is all quite a mouthful, let me provide you with an image: imagine the mesmerizing flight of a murmuration of starlings. Seemingly random patterns emerging mid-flight, which, upon closer inspection, present themselves as the intricate dance of each bird’s inert understanding of their “North Star” (their migratory path), a constant evaluation of the environment, and an orientation and coordination of the other birds in close proximity. No CEO, no org chart, no five-year plan – just a bunch of birds moving gracefully from A to B.
We advocate for organizations to think of themselves much more like this flock of birds rather than a rigid org chart with neatly drawn lines and boxes.
Dee Hock himself remarked: “The essential thing to remember, however, is not that we became a world of expert managers, but that the nature of our expertise became the creation and control of constants, uniformity, and efficiency, while our need has now become the understanding and coordination of variability, complexity, and effectiveness.”
The other day, for our Disrupt Disruption podcast, I spoke to my (and Jane’s – you might see a pattern here!) former boss, Scot Wingo. Scot was the cofounder of ChannelAdvisor, an omni-channel ecommerce company which he IPO’d at the NYSE. He now works on his next venture (ReFiBuy.ai) – an AI-native company in the ecommerce space, tackling a pesky problem vexing merchants all around the world.
In our conversation (the corresponding podcast will drop sometime next week), Scot made a comment that made me stop dead in my tracks: We talked about AI agents—those semi-autonomous pieces of software which, with the help of LLMs, act on your behalf—and how you, as the human, manage them. His response was: “We needed to give it some structure, so we used typical organizational management methods. We created an organizational chart and discussed it with the boss (boss is Scot’s term for an AI agent managing other AI agents). He then communicated the tasks to the team, which has clearly defined roles within the structure.”
Boom! And before you know it, the dreaded org chart is back (and back with a vengeance). In a world of (potentially) millions of AI bots, with many hundreds, thousands, or even millions working for you, the way you manage them is… not as a chaord, but in a rigid structure with clearly defined roles, rules, and responsibilities.
I firmly believe that chaordic structures are vastly superior for any organization grappling with an uncertain and ambiguous environment, as they allow for a fertile environment. Here is Reed Hastings, founder of Netflix, on the topic:
“The nonintuitive thing is that it is better to manage chaotically if it’s productive and fertile. Think of the standard model as clear, efficient, sanitary, sterile. Our model is messy, chaotic, and fertile. In the long term, fertile will beat sterile.”
When it comes to machines, it is time to dust off the good old hierarchical management handbooks and start writing those job descriptions.
@Pascal