Here is what is actually happening. Your buyers are not making most of your inventory decisions. Your system is.

Walk through a normal day. Every reorder point in your ERP is a decision about when to buy. Every min and max is a decision about how much to carry. Every safety stock level is a decision about how much risk to absorb. Every suggested purchase order that lands in a buyer's queue is the result of dozens of those decisions stacked together and executed automatically.

Across forty thousand SKUs, that is not dozens of decisions a day. It is thousands. And almost all of them happen without a human pausing to consider whether the assumptions behind them still hold.

The buyer enters the picture only when something looks wrong. A stockout on an A-item. An obvious overstock someone trips over in the warehouse. A number that's far enough off that it catches the eye. The rest of the time, the system decides, the system executes, and everyone assumes the buyer is managing inventory.

The system is not really deciding anything. It is executing decisions you already made.

Inventory is simply where this becomes impossible to ignore, because the volume is so high that no one could review it by hand even if they tried. You are not running an operation where people make the decisions, and the system keeps the records. You are running one where the system makes the decisions and people occasionally check the records.

But here is the part worth slowing down on. The system is not really deciding anything. It is executing decisions you already made. Every reorder point, every safety stock level, every replenishment rule was a management judgment at some point. It was a deliberate call about how to balance availability against the cost of carrying it. Someone made that call, and it was probably right when they made it. Then it got encoded into the system, and the system has been carrying it out faithfully ever since.

That is what the ERP actually does. It does not exercise judgment. It executes yesterday's judgment, exactly as given, the same way every day, long after the conditions that justified it have changed. No one consciously decides to keep carrying eight weeks of a part. The decision was made once, reasonably, and then nobody went back to it. The system did not get it wrong. It got it right, once, and never noticed when right turned into wrong.

Once you see it that way, the everyday problems start to look different.

Lead times stretch during a disruption, so the system's reorder points climb to protect availability. That is the system doing its job. Then the lead times normalize. But nobody resets the reorder points, because nothing looks wrong. The shelves are full. Service levels are fine. So the system keeps buying to a lead time that ended a year ago, and the excess sits there as trapped cash that no report flags, because by every measure the system is told to watch, everything is working.

A large customer leaves. The revenue hit shows up in the financials immediately and everyone feels it. But the safety stock that was built around that customer's demand pattern does not adjust on its own. The system is still protecting against variability that walked out the door with the account. Months later you are still carrying inventory for a customer you no longer have, and it never occurred to anyone to look, because losing the customer felt like a sales problem, not an inventory one.

None of this is a buyer error. In every case, the person did exactly what the job asks. The problem is upstream of the person; in a layer of decisions nobody is responsible for governing.

This is the heart of it. Most distributors have real governance for the decisions people make. There is a pricing authority schedule. There are purchasing approval thresholds. There are limits on what a branch can commit without a sign-off. We have spent decades building structure around human decisions.

Almost no one has built the equivalent structure around the decisions the system makes automatically. The reorder logic, the safety stock formulas, the replenishment parameters. Together, those assumptions drive thousands of daily decisions. They were set once, often during an implementation years ago, and they have been running ever since on autopilot.

We govern the decisions we can see a person making. We do not govern the decisions we made once and handed to the system to carry out. Those are still our decisions. The system is simply where they live now. They are executing every day, at scale, on conditions that no longer match the ones we decided under, and no one owns the job of checking whether they still hold.

The distributors who get ahead of this build a capability most never name. They develop the confidence that their automated decisions still reflect today's business without someone having to catch each one by hand. That capability is what I call Operational Trust.

It is the same capability that lets a leader step out of a routine pricing call, applied to the decisions a system makes instead of the decisions a person makes. Most distributors have built some version of it for their people. Very few have built it for their systems, even though the systems are now making the larger share of the decisions.

So, here is the question worth carrying back to your next replenishment review.

When your inventory system recommends a purchase order tomorrow morning, how confident are you that the assumptions behind that recommendation still reflect today's business — and not last year's? Not whether the number looks reasonable. Whether the logic underneath it has been governed by anyone, recently, on purpose.

If you are not sure, you have found the decisions in your business that no one is actually managing.