Five years ago, I wrote about remodelling my tennis serve. With the help of my coach, I learnt that commitment and around 20 hours of intelligent practice could produce meaningful progress, even when the habit being changed had been reinforced over decades.

The coach was pivotal. He saw what I couldn’t, offered expert technical guidance at the right moment and challenged my optimism. Was I merely hoping to change, he asked, or was I truly committed?

Five years on, things are very different. My serve is far more effective and powerful, and I’ve been selected to represent my county.

The world of learning also looks very different. We are seeing AI play a rapidly growing role in learning and development, giving leaders new ways to access knowledge, reflect on challenges and support their practice between coaching conversations. Generative AI can provide instant feedback, explain complex ideas, suggest ways to practise and give us access to knowledge that previously required hours of searching, or an expert sitting alongside us.

So, do we still need the 20 hours? And do we still need the coach? I think we do. Perhaps more than ever.

Information is not transformation

AI can accelerate the early stages of learning. It can help us understand a concept, identify options and notice patterns we might otherwise miss. Used well, it can make practice more focused and feedback more immediate.

But understanding what to do has rarely been the main obstacle to change. Most leaders already know they should listen more carefully, have difficult conversations earlier or create greater clarity. Tennis players often know they should relax, move their feet or watch the ball.

Knowing is not the same as doing, particularly under pressure. The real challenge is often understanding what prevents us from changing, and then staying engaged with the work when familiar habits reassert themselves. This is where a coach brings something fundamentally different: an understanding of the human being, not simply the problem to be solved.

Walking alongside, not instructing

Changing a familiar pattern often feels uncomfortable before it feels effective. A remodelled serve can initially feel awkward and unreliable. A leader behaving differently may feel exposed, less fluent or uncertain about how others will respond.

AI can help us anticipate these reactions, but it cannot experience them with us. We must still tolerate the discomfort of being a beginner, loosen our attachment to familiar ways of operating and continue when results are disappointing. That emotional work is part of learning, not an interruption to it.

A good coach walks alongside the learner through this unfamiliar territory. They help make sense of what is happening, notice patterns that are difficult to see from the inside and judge when to challenge, when to encourage and when to create space for the learner to find their own answer.

In leadership development, the goals must belong to the leader. Once chosen, however, a coach can hold the leader accountable to them, particularly when daily pressures make old habits easier and more attractive. AI can generate questions, but a coach can judge which question matters now. AI can suggest a plan, but a coach can notice when it is being avoided. AI can offer feedback, but a coach can understand the person, the context and the emotions beneath the behaviour. This is why the relationship is not simply the vehicle for development. It is central to the development itself.

Habits still need practice

Lasting improvement requires intelligent practice: focused attention, good feedback, appropriate challenge and time to absorb what we are learning. It also needs purpose. Why does this change matter? What might become possible? Are we committed, or simply hoping improvement will happen?

Desire provides energy. Purpose gives the work meaning. Intention directs attention. Practice gradually turns possibility into a reliable habit. Development does not happen at the moment of insight. It happens when insight is translated into repeated behaviour, especially in the situations where changing is most difficult. The coach helps sustain that process and AI can support it, but neither can practise on our behalf.

Making the hours count

AI will not make development effortless. It can, however, make our effort more intelligent. We can arrive at coaching conversations better informed, recognise patterns sooner and use the time between sessions more purposefully. AI can extend and enrich the development process, while the coach provides the relationship, judgement and accountability that help turn learning into change.

But we must still practise, work through doubt and vulnerability, and repeatedly choose to behave differently. Whether developing a serve or developing as a leader, there is no shortcut around the human work of change.

The question is not whether AI can save us the 20 hours. It is how the learner, the coach and the technology can make those hours count.