Calculating your OEE without a costly MES: AI for industrial SMEs
In most workshops I visit, OEE already exists. It’s read by hand by the team, jotted on a sheet pinned near the machine, then re-entered the next day into a spreadsheet no one really opens. The data is there, the capture effort is done, and yet it serves nothing. It’s the classic waste of an industrial SME: we measure, but we don’t use it.
The MES isn’t the only way in
When it comes to managing overall equipment effectiveness, the reflex is to think of a full MES. It’s a heavy, long project, often out of reach for an SME’s budget. So many give up on any visibility and stay on paper. Yet between the abandoned spreadsheet and the six-figure MES, there’s a huge space we overlook. You don’t need to instrument the whole plant to start seeing clearly into your stoppages, your micro-breaks and your speed losses.
Consolidate the readings, without reinventing everything
What AI does very concretely is consolidate what’s already captured. The morning, afternoon and night readings, across different machines and inconsistent formats, can be gathered automatically instead of recopied by hand. From there, you build a dashboard you modify in plain language: you ask for a per-line view, you add a breakdown by stoppage cause, you change the period, simply by saying it. No code, no spec frozen two years in advance, no report you no longer dare touch for fear of breaking everything. You adjust at the pace of the workshop.
The honest prerequisite: clean first
Here I have to be straight, because it’s the mistake I see most often. If your readings are dirty (stoppages badly categorized, rough durations, a machine logged as stopped while it was running), no AI will save that. Putting intelligence on false data only amplifies the false, faster and with more confidence. So we start by making the capture reliable and cleaning the history. It isn’t the glamorous part of the project, but it’s the part that decides the outcome. Skipping this step buys you a nice dashboard that lies.
The payoff is measured in weeks
The appeal of this approach is speed. The data already exists, the hardware investment is low, and the team sees a first usable dashboard in a few weeks, not in several months. It’s exactly the kind of first step that creates internal proof: you finally identify the two or three stoppage causes that truly drag OEE down, you act on them, and you make people want to go further. The MES may come one day, when the need is mature and the budget justified. In the meantime, you don’t stay blind just because you can’t afford the big tool.
For the full picture, read the guide AI in industry. See also: The data that sleeps. Wondering where to start? Gauge your AI maturity in 2 minutes, or let’s talk for 20 minutes.