AI can provide an answer in seconds. But will an employee still feel comfortable asking a colleague what might seem like a stupid question?
AI can formulate feedback. But will the manager still have the difficult conversation afterwards? AI can generate ten possible solutions.
But will someone on the team still say:
“I think we might all be wrong here.”?
Perhaps this is one of the most underestimated challenges of the new world of work. We talk about AI, productivity, automation and new skills.
Yet the more powerful technology becomes, the more important one surprisingly human resource may become:
Trust.
A team can have the best tools in the world. But if people withhold information, hide mistakes or are afraid to challenge one another, much of that potential will remain untapped.
AI is changing the way we work. Trust determines what we make of it.
Companies are investing in AI tools, agents, automation and training. But technology never enters a vacuum. It is introduced into existing teams.
With existing relationships. Existing conflicts. Existing leadership structures. And existing trust — or mistrust.
That is why one of the key questions is not simply:
“How good is our technology?”
But rather:
“How well does the human system in which we use it actually work?”
An unproductive meeting can be summarised perfectly. It is still an unproductive meeting. AI can help prepare a bad decision faster. It is still a bad decision.
And a team in which nobody challenges anyone else does not automatically become better simply because each individual works more efficiently.
Psychological safety is becoming a key skill for the future
Trust within a team does not mean that everyone always agrees. Quite the opposite. A strong team needs people who feel able to say:
“I don’t know.”
“I need help.”
“I’m not convinced by this answer.”
“I made a mistake.”
“Your idea is better.”
That requires psychological safety: an environment in which people can ask questions, show uncertainty and challenge others without fearing negative personal consequences. This becomes particularly relevant when working with AI.
Because learning means experimenting. And experimenting means that sometimes things will not work. People who are afraid to reveal their mistakes are more likely to hide poor results than to learn from them together.
So the question is not only:
“How do we get people to use AI?”
But also:
“Are we creating an environment in which people can learn together?”
What happens when efficiency replaces human interaction?
AI can make many small interactions unnecessary. In the past, a junior employee might ask:
“Could you take a quick look at this?”
Today, they ask a copilot. In the past, two colleagues might discuss an initial idea together. Today, AI can generate ten variations. That is efficient.
But those small conversations served another purpose as well. They helped people understand how others think. Who to ask about a particular problem. Who stays calm when something goes wrong. Who enjoys sharing knowledge. Who asks good questions.
That is how relationship capital is built.
Not in one big moment, but through hundreds of small ones. The danger, therefore, is not that AI will fundamentally make communication worse.
The danger is that companies may optimise away the very human interactions that would have built trust over time.
Trust cannot be taught through a PowerPoint presentation
Companies can talk about trust. They can define values. Run leadership training programmes. Offer workshops on psychological safety. All of that can be useful. But trust itself develops differently.
Through experience.
I share an idea I am unsure about — and I am taken seriously.
I make a mistake — and we solve the problem together. I need help — and someone supports me.
I hand over responsibility — and my colleague takes ownership. Trust develops through many small pieces of evidence that show us collaboration works.
And that is precisely why shared experiences are gaining new significance in an increasingly digital workplace.
Why team building may become even more important in the age of AI
A teamevent cannot repair a damaged company culture. An escape game cannot resolve a leadership conflict. a company outing is no substitute for good leadership. Nor should it promise to do any of those things.
But a good team experience can create something that may be becoming increasingly rare in the digital workplace:
a situation in which people genuinely depend on one another.
A shared problem. Incomplete information. Time pressure. Mistakes. Uncertainty. And other people whose ideas you need. In an escape game this quickly becomes visible. One person discovers a clue.
Another understands what it means. A third connects it to something that happened ten minutes earlier. Suddenly, the team needs exactly the same things that matter in everyday working life:
listening, sharing information, asking questions, making decisions and, sometimes, trusting someone else’s idea.
Nobody has to announce:
“Now we’re going to build trust.”
The situation itself creates small moments in which people experience that collaboration works.
“I had no idea that’s how you think.”
At work, we know each other through our roles.
CEO.
Team Lead.
Senior.
Junior.
Finance.
Marketing.
Sales.
In a new environment, those roles briefly become less important.
Perhaps the youngest person on the team discovers the crucial clue.
The colleague with the strongest analytical skills gets stuck.
A quiet team member recognises the pattern.
And the team lead decides to follow an idea that is completely different from their own.
Moments like these can change the way people see one another.
Perhaps that is why one of the most valuable things someone can say after a shared team experience is not:
“We won.”
But:
“I had no idea that’s how you think.”
We see moments like these again and again at URBANescape. Not because we tell teams how they should work together. But because we give them a shared mission in which they simply have to.
Trust is not a soft skill. It is infrastructure.
The more operational work AI takes over, the more important the human mechanisms that cannot be automated may become.
Judgement.
Responsibility.
Communication.
Relationships.
Trust.
AI can make employees more productive.
It can make information available faster.
It can help prepare decisions.
But it cannot guarantee that people will challenge one another when necessary.
That mistakes will be addressed openly.
That someone will ask for help.
Or that someone will say:
“I trust your judgement.”
Perhaps companies should therefore take a second transformation just as seriously as the technological one:
the human transformation.
Not only:
“What can AI do for us?”
But also:
“Which human relationships must we make sure we do not accidentally optimise away?”
Because the best teams of tomorrow will probably use outstanding technology.
But their real competitive advantage may lie in how much their people trust one another.
And that is exactly why collaboration should not only be discussed.
It should be experienced together, again and again.