The Inventory You Cannot See.

It lives in a spreadsheet, a co-supplier’s inbox, and someone’s head. Midmarket plants run without a real inventory system. Rare and long-lead parts get bought twice, or not at all. The people who know what to order are stuck reconciling instead of deciding.

That is not an ERP problem. Many of these plants have no ERP or MES to begin with. The work is making stock visible, making the next purchase order obvious, and getting hard-to-get parts moving before the job stalls.

If you cannot see it, you cannot order it in time.

Here is how ordering actually works in most midmarket plants: someone opens last week’s spreadsheet. They guess what is on the shelf. They email a co-supplier about a long-lead driver, a custom housing, a chemical that only one mill still makes. Two days later procurement places a PO for parts that were already allocated to another job.

Everyone knows this is insane. No one fixes it.

Why? Because the knowledge sits in heads, not in a system. The ROI on capturing it looks smaller than a new production line. So the line gets funded. The inventory stays invisible. Rare parts stay late.

The models exist. They cannot order what they cannot see. Strategy decks do not place a PO for a long-lead part.

The consulting-industrial complex has sold manufacturers on “AI strategy” for years. Endless assessments. Roadmaps to nowhere. Pilots that never scale. None of that tells you whether to reorder the rare driver this week.

You cannot query inventory that does not exist in queryable form. You cannot decide what to order if the count lives in a constructor’s notebook and a supplier thread from Tuesday.

The inventory layer is the moat. Not the models.

Mid-market manufacturers are stuck in between.

They have big-company complexity with small-company budgets. Multiple lines, multiple products, customers waiting on a part that takes twelve weeks, and inventory that is too complex for tribal knowledge and too thin for a $20M ERP.

These companies spend 40 hours a week reconciling stock. They discover a rare part is missing after the job is already promised. They order twice because no one can see what is allocated.

That is where rare-parts inventory actually breaks.

Multi-year ERP programs break down. By the time they are done, the long-lead part has already slipped three jobs. The champion has left. The ROI model is obsolete.

What actually works: live inventory truth, co-supplier ordering for hard-to-get parts, and decisions in days, not months. Forward-deployed engineers who sit with the people who actually place the POs, not in a consulting firm’s back office.

Capture what is stuck in heads. Make stock, allocations, and rare-parts lead times queryable. Put the next order in front of the person who can approve it.

What’s missing: someone to assemble that for manufacturing. Deployable in weeks. Maintainable by normal people.

The factories that win in the next decade will know what they have, what is rare, and what to order next, in time to act.

Not a stock report compiled last week. They will know which long-lead part is unordered, which PO would double-buy allocated stock, and where cash is sitting in the wrong inventory, before any of it costs them.

They will not look like science fiction. Same equipment. Same operators. Same co-suppliers. The difference is invisible: an inventory layer that makes the next decision obvious.