CEB (now Gartner) found that up to 60% of new managers underperform or fail inside their first two years. If you are reading this looking for how to manage better, odds are you are not one of the ones failing. You are one of the ones paying attention, and your playbook isn’t working.

Managing used to be assigning tasks, checking work, and running the review cycle. Agents do most of that layer now. Your people ship faster than they ever did. Output is cheap. You can no longer evaluate someone by what they produce, because almost everyone produces.

What is left to evaluate is judgment. Whether the person on your team caught the subtle thing in the brief that the agent missed, or asked the second question instead of accepting the first plausible answer, or noticed when the output looked right and wasn’t. Thoughtlessness is the human equivalent of AI slop. When someone is mindful and pragmatic, the results mirror the approach. When they are not, the work looks fine until the day it doesn’t.

Evaluating that is different from what you were trained to do. You cannot grade it like a deliverable. You have to talk about decisions. Walk through why the person made the call they made. Ask what they considered and ruled out. That conversation only gives you real information if the other person feels safe having it. If your feedback is punitive, they will learn to show you only the decisions that already worked. You lose the signal on everything in between, which is most of the important stuff.

Courage to experiment comes from what happens when an experiment fails. If you make visible uncertainty expensive, you get quiet uncertainty, which is worse. If you treat “I’m not sure yet” as signal instead of weakness, you get a team that thinks out loud. Thinking out loud is the habit everything else in the agentic age depends on.

None of this is clickable. You cannot learn to give non-punitive feedback by watching a module on non-punitive feedback. You learn by rehearsing, by catching yourself in the moment you would normally react badly, and by having the reaction go differently enough times that it becomes your default.

There are no models for this yet. Everyone is making it up. The honest thing to say to your team is that you are figuring it out alongside them, and that you would rather be corrected than perform. That admission is the shift. The manager who acknowledges they are still learning creates permission for everyone else to do the same.

If you want a place to practice the hard conversations before you are in them, that is what we are building Bubot for. The performance review where the honest answer is “I don’t know yet if you’re ready.” The moment an employee takes a swing and misses. Nobody’s job is on the line. By the time you are in the real version, you have already had it three or four times.

The lessons apply on both sides. Managers have to learn to coach for judgment. The people they manage have to learn to make their judgment legible. Nobody gets there by clicking Next.

If this is where you are right now, take a look at https://www.bubotlearning.com or reach out at info@bubotlearning.com. I’d like to hear what it looks like on your side.