The Middle Manager and AI: The Conversation No One Is Having

The executive conversation about AI is happening. The frontline worker conversation about AI is happening. The conversation that is not happening with nearly enough seriousness is the one about middle managers.

Middle managers are the people through whom AI initiatives either succeed or quietly fail. They control how work gets organized, which tools their teams actually use, and whether new systems are treated as genuine improvements or additional overhead to work around. When an AI deployment struggles, the root cause is often a manager who was not included in the design process and has no particular reason to champion a system they did not ask for.

This is not a malicious dynamic. It is a structural one. Middle managers are accountable for near-term team performance. New AI systems, even good ones, introduce short-term friction. There is a period of adjustment, a learning curve, a tolerance for the system getting things slightly wrong while it is calibrated. That friction cost is borne by the manager. The long-term benefit often accrues to someone else's budget line.

The organizations navigating this well are doing a few specific things. They are including managers in problem definition, not just deployment. They are making the ROI of the initiative visible at the team level, not just the organizational level. And they are treating the manager's feedback about the system in production as a first-class input, not noise to be managed.

The AI system that looks good in a demo and fails in practice almost always has a middle manager story behind it. Find that story before you ship.