Insights · point of view
AI doesn’t just change the tools people use. It changes how work is done, how decisions are made and how teams are shaped. That is an organizational design problem — and someone has to own it.
IANIA.AI · 7 min read
The org chart most companies still run on was designed for an industrial logic: layers of people, each supervising the layer below, information flowing up and instructions flowing down. It was built to coordinate human labour. When agents start doing real work — drafting, analysing, retrieving, executing — that shape stops fitting.
The useful move is to look at work at the level of tasks, and sort them: what stays fully human (judgment, relationships, accountability), what gets augmented (a human in the lead, an agent assisting), what gets delegated (an agent in the lead, a human reviewing), and what gets automated outright. Few roles are entirely in one bucket — most are re-mixed. That re-mix, not headcount, is the real story.
As humans and agents share the work, new measures appear — the ratio of agents to people, value created per person rather than tasks completed. The old temptation is to read all of this as a headcount-cutting exercise. Companies that did so have quietly rehired: cutting people creates budget room, not capability, and quality follows the people out the door. The organizations that win redesign work to be more valuable, then staff it deliberately — human and agent.
This is the part most companies get wrong. It cannot be owned by IT alone, nor by HR alone. HR understands roles, skills, culture and change. Technology understands architecture, integration and security. Data and AI understand models, quality and evaluation. The business understands the process and the value. The agentic organization is designed by all four, together — or it is not designed at all, and simply happens to you.
Helping those four functions design it together is exactly what we do. See how →