With the passing of time, you notice how some insights stick with you and have an outsized influence on your thinking. To me, one of those insights came during my undergrad studies in psychology: The huge difference between a mere work group and a team.
I had somehow made it through childhood without team sports (or much sport of any kind) so perhaps the distinction was less intuitive to me than to others.
Unlike your average work group, I learned, a team is tightly knit. More than anything, it is marked by a quality called interdependence: All the team members depend on each other in order to get the work done. A well-functioning team must collaborate intensely.
Contrast this to the typical workgroup, where colleagues tend to work alongside each other. Tasks are divided up. You work in your swimlane. When you need to know something, you reach out to your colleague and ask. When coordination is needed, you call for a meeting. You discuss, and you go back to your swimlane.
This setup does have the benefit of being comfortable and predictable. Why take the trouble of constantly communicating, discussing back and forth, and generally just working together as intensely as a real team?
Because for some types of work, a team can produce way better results.
A well-functioning team, my I/O psychology course asserted, will often outperform a regular workgroup by a landslide if the work itself demands interdependence: When different perspectives, skills and information need to be integrated to produce the result.
It seemed obvious to me that a big share of all knowledge work fell into that bucket. Since then, I have honestly been waiting for organisations to restructure work to enable a team setup. And sure; there are industries and functions that already employ true teamwork, including Product & Engineering and multidisciplinary healthcare teams. The benefits are often striking. And yet, the lion’s share of work is still performed in traditional work groups.
Apparently, we needed something to increase the value of interdependence even further. That thing might turn out to be – surprise! – AI.
Speaking to a forward-leaning Head of People in the Stockholm tech sector the other day made this exquisitely clear.
She told me that what’s keeping her up at night is not primarily how to upskill her staff in AI, or how to survive the flood of AI-generated job applications. It’s how to organize work in the new AI reality. In the old era, she explained, each employee only had to do their part of work. She drew on her own background in athletics to explain the shift:
“Yesterday”, she said, “every employee just ran their leg of the relay, then handed the baton to the next runner. Now, everyone has to run the entire relay.”
In essence; everyone needs to be more well-versed in the entire thing you’re producing. This is for at least three reasons:
AI expands individual reach. People can suddenly create suggestions, prototypes, or fully functioning tools outside their traditional functional boundaries. In order for that to be valuable, you need to have sufficient knowledge about that area.
Human capacity moves upstream. AI automation frees up time that can be used for more value-creating work. But only if you know how value is actually created in your particular organization – i.e., if you understand more than your individual task.
AI raises the bar for problems worth solving. Part of the things you charged for yesterday can now easily be prompted or built by the customer themselves. This means you have to solve a harder problem. Those tend to cross functional boundaries.
A well-functioning team does just these things. The members engage in extensive knowledge sharing, over time making each member well-versed in the entire value chain the team is handling. If the team members hold various skillsets, and work tightly together, they are also likely to arrive at more innovative solutions than any of them would on their own.
There is, of course, another possibility: AI could simply turn more of us into highly autonomous generalists. If one person can suddenly do what previously required five specialists, perhaps we need less teamwork, not more.
I suspect the opposite will happen in much high-value knowledge work. As the cost of execution falls, organizations will compete by taking on larger, messier and more ambitious problems – and those problems will require integrating perspectives rather than merely expanding individual scope.
Interestingly, organisations have long invested in making work groups feel more like teams: psychological safety, team building, communication training, retrospectives. But the big shift ahead for most organisations isn’t to schedule another team workshop. It’s to start redrawing actual ways of working.
That means reconsidering how tasks are divided, how expertise is combined, how much of the problem each person understands, and how much time people actually spend thinking and creating together.
The age of AI may, somewhat surprisingly, turn out to be the age of the team.