Insights
AI success starts before AI
The shops getting real value from AI are not the fastest adopters. They are the ones whose data was already honest. For most machine shops, that groundwork lives in one unglamorous place: the tool room.
26 Jul 2026 · 4 min read · Philippe, Niyamis
Everyone is racing. Almost nobody is ready
There is a lot of noise about AI on the shop floor right now, and most of it skips the boring question underneath: ready for it with what? A model is only as good as the record you point it at, and the record is where the trouble actually lives.
Sandvik Coromant, of all people, put numbers on this in their manufacturing wellness research. Only about a third of manufacturers have data flowing cleanly across the shop. Roughly a fifth cannot get at their production data at all. Another chunk have the data but cannot interpret it. That is not an AI problem. That is a foundation problem, and no amount of clever software on top fixes a number that three people wrote down three different ways.
"Before AI" is not a slogan. It is a place
The useful move is to stop treating "AI readiness" as a mood and start treating it as a location. For most small and mid-size machine shops, the messiest, least-connected data in the building is not in the ERP or the machines. It is in the tool room and the spreadsheets, whiteboards and memories orbiting it.
Follow one cutter and you already know the story: the tool room logs it, the programmer types it into a list, the setter notes the pot, the operator scrawls "running low," and the manager reconstructs the cost from an invoice weeks later. Five records, none agreeing by Friday. Feed that to an AI and you get a confident answer built on a guess.
Get the foundation honest first
Readiness is not exotic. It is one shared record that everyone trusts: what you have, where it is, what moved, and what it cost. Capture every take, return and refill at the source, in the place the work happens, and the number stops drifting. Do that and two things follow. The obvious wins show up immediately, fewer stockouts, fewer duplicate buys, faster stock-takes. And the clever parts you were promised finally have something real to stand on.
That is the whole idea behind TRS. It is deliberately the layer that comes before AI. Make the tool room's data honest, then let the intelligence read it, a reorder suggestion or a question to Ask TRS answered from what actually happened, not from a dashboard's imagination.
You do not need an AI strategy to start. You need your tool room to stop lying to you. Everything worth doing with AI gets easier the moment it does.