How Not to Supervise - Seven Principles That Actually Matter

Hard-won lessons from the field on what actually works when building AI agent systems. Skip the hype, learn the patterns.

Supervision is where most organizational dysfunction starts. Not from malice, but from managers who treat people like interchangeable parts, demand superhuman effort as baseline, and wonder why their teams burn out or leave.

The problem isn’t that supervisors don’t care. It’s that they operate on autopilot, following patterns they inherited without questioning whether those patterns actually work. They do the same things that frustrated them when they weren’t in charge, then act surprised when their teams respond the same way they did.

Most supervision training focuses on tactics—how to run meetings, give feedback, set goals. All useful, but missing the point. The real failures happen at a more fundamental level, in how supervisors think about the humans they’re responsible for.

You can master every management framework and still fail if you treat your team like resources to be optimized rather than people with individual contexts, capabilities, and limits. The principles that follow aren’t about process. They’re about the mindset shifts that separate supervisors who retain talent from those who burn through it.

People aren’t uniform. What motivates one person demotivates another. The feedback that helps one team member improve will shut down someone else completely. Treating everyone identically isn’t fair—it’s lazy.

Your job is to understand individual contexts. One person thrives with autonomy, another needs structure. One wants aggressive stretch goals, another is already dealing with challenges you don’t see. Generalizing means you’ll get it wrong for most of your team most of the time.

This doesn’t mean favoritism. It means adapting your approach to actual humans instead of managing to an imaginary average employee who doesn’t exist.