If your first team is AI, how will you learn to lead people?

During a recent webinar I led on AI and the future of work, someone in the audience raised a question that I find particularly interesting after more than 20 years working in people and leadership development.

How will we develop the next generation of leaders if their first teams are AI agents?

Microsoft's 2025 Work Trend Index introduced the concept of the agent boss: someone who builds, delegates to and manages AI agents to accomplish increasingly complex work.

The research drew on survey responses from 31,000 workers across 31 countries. The three figures below come from the leaders in that survey and describe what they expect within the next five years:

These are expectations, rather than forecasts of actual adoption. Nevertheless, they suggest a significant change in how work could be organised.

Imagine a young professional coordinating several AI agents. One conducts research, another analyses information, and a third prepares recommendations. This person defines the objectives, delegates tasks, reviews outputs and makes decisions.

They could gain responsibility for work that previously required a much larger team.

That is an exciting development. But does it also prepare them to lead people?

Managing AI agents requires several capabilities that are valuable in leadership:

These capabilities matter whether we are working with technology or people.

However, there is an important difference.

If an AI agent produces disappointing work, we can revise the instructions, change the process or use another tool. When a person produces disappointing work, a good leader needs to understand why.

Was the objective unclear? Did the person lack experience? Were there competing priorities? Was something happening that affected their performance?

The leader needs to have a conversation, listen, provide feedback and agree on a way forward. And the experience matters to both people.

An AI agent does not need encouragement after a difficult week, an honest conversation about its career ambitions or support in rebuilding confidence after a mistake.

Managing AI agents can develop our ability to organise work. Leading people also requires us to understand people.

The question is how we ensure future leaders get sufficient experience of both.

The Center for Creative Leadership has spent decades researching how leaders develop. Its well-known 70-20-10 framework describes three important sources of learning. The proportions are a guideline, but the central finding is that leadership development depends heavily on experience.

Importantly, not all experiences teach the same lessons. Consider two project assignments.

Both assignments can be valuable. But they develop different capabilities.

Traditionally, many future leaders gained their first leadership experience by coordinating projects, influencing colleagues or supporting less experienced team members.

As AI takes over more of the coordination and execution involved in those projects, we need to consider which developmental opportunities might change or disappear.

We should be more deliberate about the experiences through which people learn to lead.

I would start with three practical changes.

1. Review development opportunities, not just tasks

When introducing AI into a team, organisations typically identify activities that could be automated or completed more efficiently.

I would add another question: what is the development outcome of doing those activities?

For example, preparing a presentation for senior management would previously involve gathering information, challenging assumptions, working with stakeholders and developing a recommendation.

AI may now complete much of the research and preparation.

That creates an opportunity to develop critical thinking skills, and to challenge or identify information sources. In other words, to learn not to trust AI blindly, as the accountability always stays with the presenter.

2. Give emerging leaders responsibility for people-related challenges

We do not need to wait until someone becomes a line manager to develop their human leadership capabilities.

Organisations can create opportunities for people to:

• Facilitate meetings where participants have different perspectives.
• Lead cross-functional initiatives without formal authority.
• Mentor or support less experienced colleagues.
• Give and receive constructive feedback.
• Help resolve disagreements and build agreement around decisions.

These experiences require people to navigate real relationships and understand how their behaviour affects others.

AI can help someone prepare for these situations, but it cannot replace the experience of having the conversation.

3. Teach people to recognise the difference between managing work and leading people

Consider a manager introducing AI agents into their team. There are two conversations worth having.

A manager who concentrates exclusively on the first conversation may improve efficiency while overlooking important development needs.

This is particularly relevant because Microsoft's research suggests AI may allow people to take on more complex work earlier in their careers. If that happens, leadership development should evolve alongside those responsibilities.

Here is something I would encourage managers to try during their next team or development discussion.

Choose one project where AI is already being used, or is likely to be introduced. Ask three questions:

The purpose is to make AI adoption part of a conversation about people's development, rather than treating the two as separate initiatives.

The World Economic Forum's Future of Jobs Report 2025, based on responses from more than 1,000 global employers, provides an interesting perspective.

It suggests a future in which technical and human capabilities need to develop together.

For organisations, the challenge is to create opportunities for both. For individuals, it is worth considering whether their professional development includes enough experience of working with people, particularly when technology makes it increasingly possible to work independently.