For a long time, organisations have often been imagined like machines. Put the right structures in place, define the process carefully enough, measure what happens and adjust the controls until the desired result appears.
There are good reasons for this way of thinking. Industrial production depends on standardisation, repeatability and reliable processes. When thousands of identical products move through a production line, precision and control matter enormously.
They still do. But much of modern work is less predictable. It depends on expertise, judgment, collaboration and people making decisions in situations that cannot be fully specified in advance. The machine metaphor becomes less useful when the work itself depends on human beings responding intelligently to changing circumstances.
Leadership may therefore require a different image.
A leader is less like an engineer controlling a machine and more like a gardener tending a garden.
You cannot order the harvest
A gardener wants results. The purpose is not simply to create a pleasant environment in which plants can enjoy themselves. Something should eventually grow.
But gardeners also understand that growth cannot be commanded directly. You cannot instruct a tomato plant to produce twelve tomatoes by Friday because the quarterly target has been increased.
What you can do is influence the conditions.
You can make sure there is enough water and light. You can improve the soil, provide nutrients, create space, remove weeds and protect the plants when conditions become difficult. You observe what is happening and intervene when necessary.
Leadership works in a similar way.
A manager cannot personally manufacture the performance of a team. The people in the team do the work, solve the problems, speak with the clients, develop the products and make hundreds of smaller decisions every day.
The leader’s contribution lies partly in shaping the environment in which this happens. Are priorities clear enough for people to make sensible choices? Do they have the information and resources they need? Do they know where they can decide for themselves and where they should involve others? Are conflicting demands resolved, or simply passed downwards? Can people raise problems before they become crises?
These questions are less visible than a KPI dashboard, but they often determine what that dashboard will eventually show.
This becomes particularly relevant as AI enters more workplaces. AI can make reporting faster, generate more analysis and give managers access to more information about what is happening. As I argued in the previous article, that may reduce some of the information-processing work traditionally associated with management.2
That can also create a temptation to manage more simply because we can measure more, even when the additional intervention adds little.
Conditions matter more than instructions
The idea that performance depends partly on conditions is hardly new.
Decades ago, medical sociologist Aaron Antonovsky developed the concept of a Sense of Coherence as part of his work on salutogenesis – the study of factors that support health rather than focusing only on disease. He described three dimensions: comprehensibility, manageability and meaningfulness.1
Antonovsky was writing about health, not leadership, but the parallel is useful. Applied cautiously to the workplace, his framework suggests that people are better able to deal with demands when they can make sense of what is happening, believe that adequate resources are available and see the effort as worth making.
Those are remarkably practical considerations for leaders.
Does the team understand what is happening and what is expected? Are the demands realistically manageable with the available resources, authority and capability? Can people see why the work matters and how it connects to something beyond completing the next task?
Leaders cannot guarantee these conditions, but they can influence them considerably.
This is also why seemingly small managerial decisions can have disproportionate effects. A leader who resolves conflicting priorities may do more for performance than one who introduces another reporting requirement. Removing an unnecessary approval step may create more capacity than asking everyone to work faster. Clarifying who is allowed to make a decision can sometimes help more than providing another detailed instruction.
Good leadership often changes the conditions around the work rather than the work itself.
The limits of control
The gardening metaphor also points to something that management systems sometimes encourage us to forget: a great deal remains outside our control.
A gardener cannot control the weather. Leaders cannot control economic downturns, geopolitical events, sudden regulatory changes or the availability of talent in the labour market. Nor can they completely control the people working for them. Employees have different abilities, ambitions, motivations and circumstances. They will not respond identically to the same situation.
More data may describe this uncertainty more accurately, but it does not make it disappear. AI can identify patterns that previously remained hidden and provide better information for decisions; it still cannot abolish uncertainty simply by analysing it more quickly.2
There is a risk that managers respond to uncertainty by increasing control.
Performance falls, so another KPI is introduced. A project is delayed, so reporting becomes more frequent. A team is under pressure, so another status meeting appears in the calendar. People are reluctant to take ownership, so another approval stage is added to make sure nothing goes wrong.
Each intervention may be understandable in isolation. Together, they can gradually make the environment harder to work in.
The manager is trying to improve the harvest while repeatedly walking across the flowerbeds.
More pressure does not automatically create more performance
This becomes particularly visible when organisations are under pressure themselves.
Targets increase. Resources become tighter. Management wants greater visibility because uncertainty is uncomfortable. Teams are asked to become more efficient, more productive and faster, often simultaneously.
Pressure can produce results. Deadlines can focus attention, constraints can stimulate creativity, and organisations are entitled to expect performance. The difficulty begins when pressure becomes the default response to every performance problem.
At some point, asking the same system to produce more with progressively less begins to affect the conditions that made good performance possible in the first place. People spend more time reporting and less time doing. Short-term targets crowd out longer-term improvements. Experienced employees become permanent escalation points because nobody has enough space to develop others. Decisions become slower because more stakeholders want reassurance before anything moves.
Managers may not be able to remove all of these effects, but good leadership starts with recognising them.
Creating good conditions is not the same as making work comfortable. Gardeners prune, and leaders also have to challenge, prioritise and address poor performance. The point is not to avoid demands, but to create an environment in which people have a reasonable chance of meeting them.
Sometimes the soil is poor
Ever tried to grow a mango tree in the Arctic?
Even the best gardener will eventually run into conditions they cannot overcome.
Leadership literature can easily place too much responsibility on the individual manager. If the team performs well, we praise the leader. If people are disengaged, collaboration is poor or performance falls, we ask what the leader should do differently.
That is often a useful question, but not every performance problem can reasonably be traced back to the person leading the team.
A strong leader cannot indefinitely compensate for an incoherent strategy, chronic understaffing or incentives that reward the wrong behaviour. A manager cannot give people meaningful autonomy if every significant decision must be approved several levels higher. They cannot create clarity when priorities change every few days without anything being removed. And they cannot protect a team forever from demands that are built into the organisation’s operating model.
Sometimes the soil is poor. Sometimes the climate is wrong. Sometimes the organisation itself needs attention.
That does not reduce the importance of leadership. It defines its boundaries more realistically.
A good gardener works with the conditions that exist, improves what can be improved and recognises what cannot be solved by watering more enthusiastically.
A more realistic view of leadership
That, for me, is the value of the gardening metaphor. It keeps ambition and limits in the same frame. Gardeners observe, decide, intervene and adapt, while recognising that they are working with a living system rather than commanding an output.
Leadership works in much the same way. It cannot guarantee performance, but it can improve the chances of it by creating clarity, ensuring that resources are adequate, placing responsibility where it belongs and giving people enough room to exercise judgment. As AI makes information, analysis and monitoring easier, this distinction may become even more important: greater visibility should not automatically lead to greater intervention.2
Targets, deadlines and results remain. The leader’s task is to shape the conditions in which people have a realistic chance of meeting them well – and to recognise when the problem lies beyond the team, or beyond leadership itself.
References
1. Aaron Antonovsky, Unraveling the Mystery of Health: How People Manage Stress and Stay Well, Jossey-Bass, San Francisco, 1987. Google Books
2. Susanne Tafvelin, Maria Forsgren and Andreas Stenling, “Redesigning the Leader Role for the Age of Artificial Intelligence: A Work Design Challenge”, The Journal of Applied Behavioral Science, first published online 20 September 2026. https://doi.org/10.1177/00218863261488834
