An employee comes to their manager excited.
They've figured something out. A task that normally takes four hours can now be done in 30 minutes using AI.
The result looks good. The employee sees an obvious opportunity. Maybe they even think the whole team should start doing it this way.
For an organization investing in AI adoption, this sounds like success.
And it might be.
But it also creates a new question:
What should the manager do next?
Should they encourage the employee to keep going? Tell the rest of the team about it? Ask them to stop until the process can be evaluated? Find out exactly how they did it? Bring someone else into the conversation?
The answer may depend entirely on the situation. And that's the point.
As employees begin experimenting with AI, managers are encountering a new category of leadership moments, often without a clear playbook for navigating them.
AI adoption doesn't stop when people start using AI
A lot of organizational AI initiatives understandably focus on adoption.
Get employees access to the tools.
Teach them how to use them.
Develop policies.
Identify use cases.
Encourage experimentation.
Measure productivity.
But eventually, something important happens:
Employees actually start experimenting.
And experimentation is inherently less predictable than implementation.
People closest to the work begin discovering applications leaders hadn't considered.
They combine tools in unexpected ways.
They question processes that have existed for years.
They find shortcuts.
Some experiments fail.
Some work remarkably well.
Some produce promising results while simultaneously raising questions the organization hasn't answered yet.
At that point, AI adoption stops being only a technology initiative.
It becomes a management challenge.
Managers may not be the AI experts in the room
There's another wrinkle.
The employee bringing the idea to their manager may understand the AI workflow better than the manager does.
That's not unusual.
AI tools are changing quickly, and employees are experimenting at different speeds.
The old assumption that the manager has more experience with the way work gets done may not always hold.
But the manager is still responsible for leading. That means the job isn't necessarily to know more about the technology. It's to know how to navigate the situation.
What problem is the employee trying to solve?
What changed?
What does the new process improve?
What assumptions are being made?
What questions still need to be answered?
Is this something the manager can green-light, or does someone else need to be involved?
Leadership in this environment doesn't always mean having the answer. Sometimes it means knowing which questions need to be asked next.
The goal can't be to shut down experimentation
Organizations want people experimenting with AI for a reason.
Some of the most valuable use cases may emerge from employees who understand a workflow deeply enough to recognize where AI could improve it.
If every new idea is met with hesitation, employees learn quickly.
Don't bother.
Stick with the old process.
Experimentation isn't actually welcome here.
That's not exactly a recipe for successful adoption.
But the opposite extreme creates problems too.
Not every AI experiment should automatically become the new way of working simply because it produced a good result once.
There may be questions about accuracy, repeatability, privacy, governance, required review, or how the workflow behaves under different circumstances.
That's the leadership tension.
How do you create permission to experiment without turning experimentation into automatic approval?
The Practice Ground editorial strategy deliberately places this tension within Leading Through AI Change: managers need to navigate experimentation alongside governance, accountability, uncertainty, and evidence—not simply teach people how to use the technology.
“It worked” is the beginning of the conversation
Imagine an employee demonstrates an AI workflow that dramatically reduces the time required to complete a task.
That's useful evidence.
But it isn't necessarily the end of the evaluation.
A manager may need to understand what “worked” actually means.
Did it save time?
Was the output accurate?
Was there rework later?
Would it work across different situations?
What information was used?
Did the workflow introduce any new considerations?
Is the employee proposing an experiment, or a permanent process change?
Those questions aren't intended to kill the idea.
They're how an organization learns from it.
And that distinction matters.
A manager who responds with an immediate “no” may shut down useful experimentation.
A manager who responds with an immediate “yes” may move faster than the organization is ready for.
In many cases, good leadership happens somewhere in between.
AI experimentation needs managers who can lead ambiguity
Organizations can create policies. They can establish governance structures. They can define approved tools and workflows.
All of that is important.
But no policy will anticipate every idea employees have as AI becomes more deeply embedded in work.
There will be gray areas. There will be new use cases. There will be situations where the result looks promising but the path forward isn't immediately clear.
And someone will have to navigate them.
Increasingly, that person will be a manager.
They'll need to encourage initiative while maintaining appropriate boundaries. They'll need to acknowledge what they don't know. They'll need to recognize when an experiment deserves further exploration. They'll also need to recognize when a question belongs with someone else.
That's not AI expertise.
It's leadership judgment in an AI-enabled workplace.
Are we training managers for that?
This is where organizations should think beyond traditional AI training.
Employees absolutely need to understand how to use AI responsibly and effectively.
But managers have another job.
They have to lead the people using it.
And that creates conversations no prompting course can fully prepare them for.
An employee discovers an unexpected use case.
Someone wants permission to try something new.
A workflow works well but raises unanswered questions.
Two employees disagree about whether an AI process is actually better.
An executive wants adoption to move faster.
A team member wants to know where the boundaries are.
These aren't hypothetical edge cases.
They're the kinds of human leadership moments that emerge as AI adoption moves from strategy into everyday work.
That's what managers need to practice
Knowing that managers should “encourage responsible experimentation” is useful.
Actually doing it is harder.
Because eventually there's another person in the conversation.
They're excited about their idea.
They've invested time in it.
They believe it works.
Maybe they're expecting their manager to be impressed.
And the manager has to respond in real time.
This is where practice matters.
Not because there's one perfect response managers need to memorize.
But because they need experience navigating the tension.
Asking questions.
Making judgment calls.
Responding to what happens next.
Reflecting.
And trying again.
That's the purpose of Leading Through AI Change from Practice Ground.
It isn't prompting training or product training.
It's practice for the human leadership moments created by AI adoption, the broader portfolio promise laid out for Practice Ground's AI-change offering.
Your employees are experimenting. Prepare managers to lead what comes next.
The question for organizations is quickly changing.
It's no longer simply: How do we get people to start using AI?
It's also: What happens once they do?
Employees will experiment. They'll find better ways of working. They'll challenge existing processes. They'll make mistakes. And they'll discover possibilities their organizations haven't considered yet.
Managers will be standing in the middle of much of that change.
They don't need to have every answer. But they do need to be ready to lead the conversation.
Practice Ground by Virbela helps managers rehearse the human leadership moments created by AI adoption before they're real.
Explore Leading Through AI Change
