The warehouse started at 7 a.m. That was the deadline. Not noon, not whenever someone got around to running the ERP report, not after the first customer called asking where their order was. Seven.
By then, the warehouse manager needed to know what to count. Customer service needed to know which orders were at risk. Inventory control needed to know where the records might be wrong. The GM needed to know what was about to get in the way of shipping.
The data was sitting in the ERP. The answers weren’t. And nobody had the time, every single day, to turn one into the other.
Data Isn't the Same as an Answer
The client was a specialty industrial equipment manufacturer running on a legacy ERP. The system held everything it needed to: open orders, open purchase orders, inventory balances, part numbers, due dates, shortages. The facts were all there.
But a fact isn’t an instruction. “This part has 12 units on hand” isn’t the same as “count this part today, because a customer order depends on it.” Turning the first into the second meant cross-referencing inventory against open orders, open POs against ship dates, and counted history against what had actually changed, for every part, every morning. That’s real analytical work, and nobody had a free hour to do it, let alone a free hour before 7 a.m.
So it didn’t get done. Not slowly, not partially. It just didn’t happen, until something forced it. A buyer could see open POs, but not which ones were actually putting an order at risk. The warehouse could see inventory, but not which counts mattered today versus any other day. Customer service could see a late order, but only after the customer had already called to ask. The GM saw all of it land on his desk at once, after the fact, as a problem instead of a warning.
Making the ERP Work Harder, Instead of Replacing It
The easy answer would have been to blame the system. The ERP was old, the reporting was weak, and the data was hard to use. That’s usually the point where companies start talking about an upgrade.
Sometimes that’s the right call. But ERP upgrades are expensive, risky, and disruptive, especially before the underlying operating process has been cleaned up. They take time, they pull good people off the business, and for a manufacturer trying to ship customer orders every day, a bad implementation can create more pain than the old system ever did.
This company didn’t need a new ERP. It needed something doing the analytical work nobody had time for, every night, automatically, so the answers were sitting there by morning instead of waiting to be assembled. No rip-and-replace. Just the missing translation step, finally getting done.
Where the Real Work Happened: Overnight, Not in a Report
The temptation with a project like this is to think the win is a nicer-looking report. It isn’t. A report that just displays ERP data faster is still asking someone to do the thinking. The actual unlock was moving that thinking itself, the cross-referencing, the prioritizing, the “does this matter today” judgment, into something that ran automatically overnight, while everyone was at home.
Copilot helped build the logic that used to live only in a person’s head: which shortages threaten which orders, which counts are stale enough to matter, which POs are tied to real risk versus background noise. Power Automate ran that logic every night and delivered the result, already finished, before the building opened. By the time the warehouse manager sat down, the analysis wasn’t pending. It was done.
That’s the difference between a faster report and an answer.
The 7 A.M. Packet
Every morning, before the floor started moving, the warehouse manager, customer service, inventory control, and the GM each got an email. It held four lists: POs to follow up on, inventory to count right now, orders at risk, and critical shortage items.
Nobody had to build those lists themselves anymore, and that’s the part that actually mattered. The warehouse manager didn’t have to wait for someone to ask for a count. Inventory control didn’t have to guess which records to trust. Customer service didn’t have to find out about risk only after the phone rang. The GM didn’t have to walk the floor piecing together fragments from four different people. The thinking had already happened overnight. The picture was just waiting in the inbox.
Count This, Not Everything
Every warehouse says it needs better inventory accuracy. True, but too broad to act on at 7 a.m. The team didn’t need a reminder to count more. They needed to know which counts mattered today, and that required real judgment: quantity on hand, quantity already allocated, outstanding POs, supplier promise dates, customer ship deadlines, and how recently the location had actually been counted.
That last piece mattered most. A location counted yesterday didn’t need counting again unless something had changed. A location nobody had checked in weeks, sitting behind a part that could delay a real order, jumped to the top.
The system didn’t say “go improve inventory accuracy.” It said: count this part, in this location, because this customer’s order may be at risk. That’s a finished decision, not a suggestion to go figure one out.
Following Up on the POs That Actually Matter
The same logic applied to purchasing. An open PO list isn’t an answer either. It’s usually too long, too flat, too easy to ignore. The buyer didn’t need everything that was open. They needed to know which POs were actually tied to risk, today.
So the overnight process checked whether current inventory and open POs were enough to hit each customer’s ship deadline, then handed back a specific instruction: expedite this PO, add quantity to that one, open a new PO where none exists, or flag the customer order that may slip. If a shortage looked likely to cause a delay, the system named the affected customer, so service could get ahead of the call instead of waiting for it. The buyer stopped chasing every vendor and started chasing the handful that could actually change the day.
A Morning Rhythm, Not Just a Report
The biggest change wasn’t the email. It was what didn’t have to happen anymore. Every morning, four people started from the same finished analysis instead of four different fragments of raw data.
Judgment still mattered. The warehouse still had to verify counts. Purchasing still had to call vendors. Customer service still had to manage the conversation. The GM still had to make tradeoffs. But the day no longer started with “go figure out what’s wrong.” It started with “act on what we already know is wrong.” That’s the real difference, and it’s the thing that used to never get done at all.
The Result
This isn’t a project with one clean before-and-after number. The win was decisions that used to never get made, now getting made automatically, every single night.
Shortages surfaced before the customer call instead of after it. Counts happened before the buyer chased the wrong vendor. At-risk orders reached the GM as a warning instead of a surprise. Customer service knew who might be affected before they picked up the phone. PO follow-up got sharp instead of scattered. And the company never had to gamble on a long, expensive ERP overhaul to get there.
The legacy ERP didn’t change overnight. What changed is that the work nobody ever had time to do got done anyway, while the lights were off. When the warehouse opened at 7 a.m., the answers were already waiting.