Drifting or leading with AI?

AI is supposed to save us time.

Often, it does. It helps us analyse information, work faster, compare options and test ideas we may not have reached alone. But AI can also turn one task into five before we have really noticed.

We open a tool with something clear in mind. One answer creates another possibility. That leads to another prompt, another comparison, one more version. None of it is wasted exactly. That is what makes it difficult to stop.

Say we open AI to prepare for a difficult team conversation. Forty minutes later, we have three versions of the message, a revised presentation and several ways to structure the meeting. What we may still not have decided is what the team actually needs from us.

AI has helped us produce more in less time. But has it helped us use that time on what matters?

AI can save time on a task while taking time from what matters more

This is easy to miss because the work feels productive. We are researching, drafting, comparing, refining, and increasingly building — an agent, a workflow, an automation. But more output does not necessarily move us closer to the decision, conversation or problem that mattered when we began.

The cost is not only the time spent prompting. It is what that time replaces: the judgement we still need to make the conversation we still need to have the problem we have not stayed with long enough to understand.

AI can make production faster without making the underlying decision easier. In some cases it does the opposite, giving us more plausible options to choose between.

The same shift happens when we move from using AI to building with it. It has never been easier to wire up an agent, automate a workflow or connect a few tools together, because it is possible. We forget to stop to check whether the process was worth making more efficient in the first place.

Activity is not direction.

Exploration is not the problem

Not every use of AI needs a precise outcome. Open exploration can reveal connections we would not have looked for, help us see a problem differently or lead somewhere we could not have planned in advance. Sometimes wandering is exactly the point.

The difference is whether we have chosen to explore openly or simply lost track of where our attention is going. If we are using AI to make a decision but are still generating alternatives an hour later, the exploration may be costing more than it gives back. If we are deliberately following possibilities to see what emerges, that same hour may be time well used.

Taking agency over our time does not mean making every minute efficient, or controlling curiosity so tightly that nothing unexpected can happen. It means being clearer about what kind of work we are doing.

Are we looking for an answer? Trying to improve something? Preparing to make a decision? Or deliberately exploring to see what we find?

All can be worthwhile. Drift begins when we think we are doing one while gradually slipping into another.

“Go and explore AI” can be an expensive instruction

The same pattern appears at team level. “Go and explore AI” can create freedom and lead to useful discoveries. It can also mean people spend hours testing, comparing and prompting without a shared sense of what the organisation is trying to understand or improve.

Leaders do not need to prescribe every step, but it helps to create enough direction around the exploration. Before beginning, agree:

• what we are exploring
• why it is worth our time
• when we will stop and review what we have learned
• what decision or next move the exploration should support

At the review point, it helps to ask whether we learned something useful, whether there is an idea worth testing in the real work, and whether we are closer to a decision — or have mainly created more options. We may decide to continue. We may also decide the exploration has done its job.

Undirected exploration still costs something. Usually time. Sometimes trust.

Disciplined curiosity helps us retain agency

Disciplined curiosity creates a little space before possibility develops its own momentum. It helps us notice what is happening, consider whether it is still useful and choose the next move more deliberately.

Three questions can help:

• What was I actually trying to achieve?
• Is this still worth the time I am giving it?
• Am I choosing to continue?

Sometimes there is genuinely more to understand, and we keep going. Other times we already have enough, and the work is to make the judgement, have the conversation or act on what we know. Either way, the difference is not exploration versus efficiency, it is intention versus being carried along.

AI will keep expanding what we can do. Leadership increasingly sits in deciding what deserves our attention, what is good enough and when it is time to move.

Where is AI helping your team move forward and where is it mainly creating more to respond to?