Information that no longer dies at the end of the shift

The problem

The line manager knows what changed. Nobody writes it down.

The line manager's hands are busy adjusting color registration or holding a pallet. Nobody asks them anything because they can't type. Information like more registration rejects since an ink batch change, or an operator needing backup, gets lost shift after shift.

• The line manager spends the shift walking between the litho press, the curing oven and forming, with hands busy adjusting color registration or steadying a pallet.
• They notice what no system records: more registration rejects since an ink batch change, slippage on side seams at the body welder, an operator who needs backup on the seamer.
• Nobody asks them anything because they cannot type, and writing it down means walking back to the office.
• So the information dies at the end of the shift, and the same issue is rediscovered by the next manager days later.

How it fits the IRIS system

Connect in voice mode — an earpiece, a wake-word and the line manager's own words.

Connect voice mode with a Bluetooth earpiece and a custom wake-word. The line manager dictates while walking; an LLM structures the dictation as an issue linked to the active line and shift, and returns support by voice or push in seconds.

It records what the line manager chooses to say, when they choose to say it, and ties it to the line and shift that were running. That is how a hunch about an ink batch becomes a pattern somebody can act on.

See the full IRIS architecture →

Before and after

The shift as it is remembered versus the shift as it is dictated

Aspect	Today	With iLEAN Connect
Register rejects after an ink batch change	Mentioned at handover, if remembered	Logged with line, shift and time
Side seam issue at the body welder	Lives in one person's head	Structured as an issue on that asset
Where it is written	Nowhere, or a notebook	Central memory, queryable
Line manager's effort	Stop, walk, type	Speak while walking the line
Answer to a question on the floor	Go to the office terminal	Back by voice or push in seconds
Talk on the floor without the wake-word	—	Ignored

Operational knowledge that dies at shift end → enters the system in real time and gets cross-checked with live data.

Impact estimate

No standalone payback — and we do not force one.

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

• This case has no standalone payback and we do not put a number on it: its value is hard to monetize upfront, and a forced figure would weaken the credibility of the rest of the matrix.
• It is an enabling piece for continuous improvement, because it stops shift-level knowledge being lost shift after shift.
• The change is qualitative but immediate: issues enter the system in real time and are crossed with live line data.
• The right way to size it is with your continuous improvement lead, counting the issues that are rediscovered today instead of solved.

Hard to monetize upfront, but it stops continuous-improvement information from being lost shift after shift. Estimate to validate with the continuous improvement lead.

And the fair question from the production manager

“Is this a way to listen in on the line?” — no. iLEAN only listens after the wake-word, only to the line manager, and only to what they decide to dictate. Turning a dictated sentence into an issue with line, process and shift is an anchored task where the best models drop below 1.5% error [1], and the manager sees the structured issue before it is saved. What they gain is not having to remember at the end of the shift what they saw eight hours earlier.

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