What happens to leadership when part of the workforce is AI agents? It looks more and more like leading a team, and it moves the human task from doing to judging.
Imagine an employee you recruit, onboard, keep training, reward for good behaviour and, in the end, let go when a better qualified candidate turns up. It sounds like the classic HR cycle for a person. In the title, AR is our shorthand for agent resources: what HR is for people, but for AI agents.
Ismail Amla, a senior vice president at Kyndryl, suggests we treat AI agents in much the same way: as full members of the workforce, who are onboarded, given a defined role, coached and eventually retired.
We are on the threshold of a reality where the workforce becomes a hybrid of people and intelligent agents, working side by side. That changes fundamentally how we understand leadership, IT and, not least, our own jobs.
When the head of IT and the head of HR move closer together
When AI goes from being a simple tool to being an autonomous agent, something happens on the organisation chart. The line between the IT department and HR starts to blur.
Leading a fleet of AI agents is strikingly similar to leading a human team. They need guidance, they need to be held to goals, and their behaviour needs fine-tuning. That is why more leaders now talk about bringing the head of IT and the head of HR closer together, because "managing IT" and "leading people" are becoming two sides of the same coin in the age of agents.
Facts
Semafor wrote in April 2026 that chief technology officers and chief HR officers have more to do together as companies take on more AI agents. The CEO of Pearson, Omar Abbosh, calls it a strong partnership, where one provides the tools and the other the culture. The CEO of Automation Anywhere, Mihir Shukla, wants the two to plan together which jobs disappear, change and emerge (Semafor, 17 April 2026). KPMG India describes the head of IT as the one who puts the agent into operation, and the head of HR as the one who brings it into the workforce, each with their own responsibility (KPMG India, 30 June 2026).
But if agents can take over the tasks, carry out the strategies and optimise the processes, what is left for us? What should people actually do?
The danger of agentic drift
The most important task for people will be to bring meaning into the loop. AI agents are extremely literal. If you ask a purchasing agent to always find the lowest price, it will do so, even if that means it drops the long-term, strategic partnerships your business has spent decades building. The example is imagined, not a real case.
We call it agentic drift. The technology delivers exactly what we ask for, but not necessarily what we actually want or need. The agent hits its measurable KPI but misses the overall purpose. Our job becomes making sure we still want what we asked for.
What should people do?
When agents take over the doing, people move up to the helicopter view. The jobs of the future are not about doing, but about steering, judging and connecting. Here are three job profiles that we think will become important.
1. AI Prompt & Behaviour Architects. Where IT developers used to write code, tomorrow's employees will design behaviour. These profiles do not just tell the agent what to do, but define its ethical limits, its rules of the game and its criteria for success. They act as a kind of "shop stewards" and culture carriers for the machines, making sure the agents act in line with the company's values and long-term strategy.
2. Context & Meaning Evaluators. This is the safeguard against agentic drift. These employees keep watching the agents' results with critical, human eyes. Their task is to ask the question: "Yes, the price came down, but did we lose our soul and our relationships along the way?" They bring intuition, empathy, market context and common sense into play, everything an algorithm cannot read from raw data.
3. Cross-Functional Orchestrators. When processes are automated, there is an enormous need for people who can build bridges. Between the different AI agents, so that they do not work against each other, and not least between people and machines. Tomorrow's leaders will have to master the discipline of leading teams where half the department drinks coffee and the other half uses electricity. It takes a deep understanding of organisation, development and co-creation.
Our take
The three profiles are our take on which tasks stay with people once the agents do the executing. They are not established job titles. The words are our own.
Questions
We do not yet know what this looks like in a small business, where IT and HR are not two departments and may sit with the same person. Who owns the agents there?
Conclusion
AI agents are not coming to make us redundant, but to force us to be more human. When routine and execution are handed over to the digital colleagues, the most important task for people becomes providing critical judgement, empathy and strategic direction. We will no longer need to be the best at finding the answers. We will need to be the best at asking the right questions and judging whether the answer makes sense.
The agent delivers what you ask for. Your job is to make sure you still want it.
Sources
- Ismail Amla, Kyndryl, in Fast Company: AI without culture change is just a failed proof of concept, 2025-12-17
- Semafor: AI Is Making Chief Tech Officers and Chief Human Resources Officers Work Together, 2026-04-17
- KPMG India: Agents are now on the org chart. Who is managing them?, 2026-06-30

Get it in your inbox
Write to us, and we will put you on the list. Tell us what you would most like to hear about: perspectives, fables or news from Coga. You unsubscribe by writing to us again.
Write to raadgiver@coga.dk