Source: Koponen et al. (2025) — “Work Characteristics for Middle Managers Leading AI Service Teams,” Journal of Service Research
Middle managers already had a difficult job.
They sit between strategy and reality, which is basically a polite way of saying they receive pressure from above and confusion from below.
Now add AI service teams.
This paper studies how middle managers experience AI integration in service work, especially in financial services. The useful point is that AI does not enter the organisation as one simple thing. Sometimes it feels like a technical tool. Sometimes it feels like a coworker. Sometimes it feels like a mysterious new team member that never sleeps and occasionally needs supervision.
That changes the manager’s work.
Managers are no longer only coordinating people. They are coordinating relationships between people and systems. They have to understand what the AI can do, what it should not do, when employees trust it too much, when they ignore it too much, and how customers experience the result.
This is a lot.
It also means middle managers become translators.
They translate strategy into workflows. They translate AI outputs into operational decisions. They translate employee concerns upward. They translate customer pain into system improvements. If that sounds exhausting, it is because it probably is.
The interesting part is that AI does not remove management work. It changes the texture of it.
Less pure supervision. More sense-making.
Less “did the employee do the task?” More “how do human and AI roles fit together without making everyone quietly resent the dashboard?”
My takeaway: AI transformation depends heavily on middle managers, even though they are rarely the stars of the AI story. If companies want AI service teams to work, they need to support the people managing the messy middle. That is where adoption becomes real.