OEE in a 24/7 plant — the three shifts perform differently, and nobody knows why until Monday.

The problem

Three shifts, three realities — and an average OEE that hides what matters.

The operations director of a 24/7 plant knows the sentence by heart: “the average OEE looks fine, but nights drag it down”. And they know the next chapter too: the Monday meeting where the night shift leader says nights always run short-staffed, the afternoon leader says the line is handed over badly loaded, and the morning leader says they break records but pick up the worst of the SKU changeover. Everyone is partly right and nobody holds the full truth, because each one measures their own way.

1. Each shift reports however it can — mornings on a screen, afternoons on paper, nights in a spreadsheet filled in at the end. The micro-stops (the ones that weigh most on real OEE) get reported in none of the three.
2. The verbal handover loses 80% of the context — the incoming shift inherits a line whose last state nobody described in full. It starts at 22:00 by working out what happened at 21:50.
3. The weak shift becomes chronic — with no comparable data, the problem gets attributed to people (“the night team is lazy”). The team then goes on the defensive, and the conversation turns sour.

The classic anti-pattern is measuring badly and arguing in the meeting. The argument does not build OEE — it builds distrust between shifts.

How it fits the IRIS system

iLEAN does not add another dashboard — it captures each shift’s reality and passes it to the next.

The IRIS category (Industrial Reality Intelligence Systems) starts from one principle: capturing reality comes first, reports come after. Applied to OEE, that means not trusting manual reporting: Edge watches the line, Brain consolidates, the agent delivers what each shift needs to see. iLEAN acts as the filler between the three shifts and the systems you already run (MES, SCADA, ERP), without asking you to change them.

Edge measures the line with the same yardstick for all three shifts. Brain consolidates without blaming anyone. The agent prepares the handover. The warm start stops being a slogan.

The three iLEAN pieces applied to OEE in a 24/7 plant:

• Edge — a terminal with vision and sensors over the line that captures cycles, micro-stops and defects in real time. Without asking the operator to fill in a screen. It works without a network — if the plant loses WiFi, Edge keeps counting cycles and logging stops, because the shift’s data cannot depend on connectivity.
• Brain — consolidates data from the three shifts under the same OEE definition (availability × performance × quality). It shows the weak shift with data, not with opinion, and breaks it down by root cause (micro-stops, changeovers, scrap), not by person.
• Agent — prepares the automatic handover report for the incoming shift: what happened in the last two hours, what is still open, the adjustments they need to know about, the spare parts consumed. And it proposes actions for the weak shift based on the root-cause pattern, not on intuition.

See the full IRIS architecture →

Before and after

Hand-reported OEE vs. captured OEE with an automatic handover

Aspect	Classic OEE across 3 shifts	With iLEAN Edge + Brain + Agent
Data source	Manual report per shift, each with its own criteria	Edge captures cycle, micro-stops and defects on the line
Yardstick	Each shift has its own	The same for all three, comparable
Micro-stops	Not reported	Counted automatically
Handover between shifts	5 verbal minutes, 80% of the context lost	Handover report ready on screen and in the earpiece
Identifying the weak shift	Argument in the Monday meeting	Weekly data with root cause, not with a culprit
Temporary staff / turnover	The veteran is the only safety net	Persistent context, delivered to whoever comes in

Impact estimate

Impact estimate for your plant — to be validated with your numbers.

The block below is an estimate to be validated against the concrete data of your plant. We put it forward so the committee has an order of magnitude; we refine it during the diagnostic.

• Continuous or discrete process plant, 24/7 across three shifts, current OEE between 55 and 70, with visible differences between shifts.
• Edge + Brain + Agent pilot on a single pilot line (the bottleneck, or the one with the largest gap between shifts). First value expected within a few weeks: as soon as Edge captures, the real OEE per shift appears and the conversation changes.
• Expected OEE improvement in the first year: +5 to +12 points on the pilot line (estimate to be validated), with the bulk of the jump coming from the weak shift levelling up.
• Indicative payback between 4 and 9 months. The hard lever is the sum of micro-losses eliminated, defects caught earlier and fewer stops caused by context lost in the handover.

And the works council’s reasonable doubt

“Is this to watch the operator?” — no. iLEAN sees the line, not the person: it is process verification, not people. The data is attributed to the shift, not to a name, and the conversation it opens is not disciplinary — it is collective improvement. What changes is that the operator no longer starts a shift asking “what did they leave me?”; they start warm, with the previous shift’s context already digested. Same person, same pay, a different working life.^[1]

[1] OpenAI paper “Why Language Models Hallucinate”, 2025 — on the reliability of AI in anchored tasks.