At 6:40 in the morning, before the first shift has found its coffee, the production planner is already doing the job nobody budgeted for. The ERP produced a schedule last night. This morning she is rebuilding it by hand — checking which work orders can actually run, which parts are really on the shelf, which supplier quietly moved a date. The spreadsheet is open. It is always open. By the time the huddle starts, the plan she presents is hers, not the system's.
The plan is right. The day is different.
This is not a software failure, exactly. The ERP is right about the plan: given the orders, the bills of materials, and the stock it was told about, the schedule it produced is correct. The spreadsheet is right about the day: the planner knows the count on aisle four is wrong, that a supplier's Thursday promise means Tuesday, that one machine prefers not to change over twice. Both are right. Nothing connects them, so a person does — every morning, from memory, under time pressure.
Four questions, every morning
Strip the rebuild down and it is four questions. What can we make? What is missing? What do we order? What is about to run out? Every production plan is an answer to those four, and every morning they have to be answered again, because yesterday's answers expired overnight: a delivery slipped, a count was wrong, a rush order landed. The planner's skill is not answering them once. It is answering them daily, against data that will not sit still.
Notice what the questions need. Not a better dashboard. A joined-up view: orders exploded through their BOMs, checked against inventory as it actually is, with supplier lead times attached — and the gaps named, with reasons, in time to act.
Shortages are found at the line because they were invisible in the plan
Ask a plant manager where shortages show up and you will rarely hear 'in the schedule.' They show up at the line, after the job has started, when the missing part stops everything and the expedite calls begin. The part was short all week. Nobody could see it, because the plan was built on a stock figure that went stale the day it was entered, and the BOM that would have exposed the shortage was never exploded against real inventory.
The same invisibility protects bad data. Duplicate SKUs. A bill of materials with a blank lead time. A unit of measure that means one thing in purchasing and another in the warehouse. In a spreadsheet these are typos. In a plan they are stoppages that have not happened yet.
Expediting is a habit, not an exception
Every plant expedites. The problem is when expediting becomes the plan: rush freight as a line item, buyers spending their mornings chasing parts instead of negotiating them, 'just in case' inventory quietly eating the warehouse. Expediting is expensive precisely because it is reactive — it pays rush prices to recover from information that arrived too late. Most of that information was in the building the whole time. It was in the ERP, in a supplier's email, in a count nobody reconciled.
What changes when the brief is already answered
The fix is not a new ERP, and it is not an AI that runs the plant. It is smaller than that: connect what already exists and answer the four questions before the shift starts. The orders, BOMs, inventory movements, and supplier lead times already live in the plant's systems and spreadsheets. Read them together, and the morning starts with a brief instead of a rebuild: these work orders can run; these three are held, and here is the exact part, quantity, and reason; this reorder needs placing today, because cover runs out before the lead time does; these BOM lines are too uncertain to plan on, and they are queued for a human to settle.
The planner still decides. She approves the purchase drafts or overrides them, and her reasons are kept, so next month's brief is smarter than this month's. What she stops doing is the archaeology — digging the truth out of four systems before she is allowed to think. Her judgment was always the scarce resource. The rebuild was spending it on reconciliation.
Plants do not need more software ceremonies. They need the truth about today, early enough to act on it. Answer the four questions before the shift, and the morning meeting stops being an apology for the schedule — and starts being a plan.