The shift lead's knowledge, captured hands-free

The problem

The person who knows why a lathe is drifting has no hand free to say so.

critical operational knowledge lives in the shift lead's head and never enters the system, because there's no way to type.

• The shift lead walks between lathes all day with hands full: a caliper, a bar end, a box of parts to check against the thread gauge.
• They know which lathe has been making more scrap since the last insert change, which parts washer is leaving residue and which new operator needs help.
• Recording any of it means walking to a terminal and typing, so it gets told verbally at handover, or not at all.
• The next shift solves the same problem again, with no memory of the previous three attempts, and the improvement team never sees the pattern.

How it fits the IRIS system

Connect in voice mode — a custom wake-word, and the lead speaks while walking.

Bluetooth earpiece + Connect voice mode with a custom wake-word. A model structures the dictation as an issue linked to the active lathe/lot/shift and returns support by voice or push within seconds.

It records what the shift lead chooses to say, when they choose to say it. That is not a detail: it is what keeps the case alive months later instead of being perceived as monitoring the floor, and it is why leads end up dictating more than anyone expected. The more it is used, the more the next shift inherits instead of rediscovering.

See the full IRIS architecture →

Before and after

The shift in someone's head versus the shift on file

Aspect	Today	With iLEAN Connect
An issue spotted on lathe 7	Remembered, maybe	Logged with lathe, lot and shift
Effort to record it	Walk to a terminal	Say it while walking
Recurring scrap on one machine	Rediscovered each shift	Visible as a pattern
Support for a new operator	Whoever is nearby	Answer by voice or push in seconds
Handover between shifts	Five minutes of talk	A written summary already there
Chat between operators	—	Never captured

knowledge that used to die at shift-end → enters the system in real time and feeds live data.

Impact estimate

An enabling piece — no payback of its own, and we say so.

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.

• No standalone payback. We do not put a number on it because its value shows up in other projects, not in this one, and forcing a figure would not survive a review.
• Every continuous-improvement effort in the plant depends on this information existing: scrap by lathe, recurring causes, the effect of an insert change.
• Knowledge that used to die at shift end enters the system in real time and feeds live data to the other cases.
• The right person to size it is your continuous-improvement lead, and we size it with them, on your own scrap and downtime history.

hard to monetize upfront, but all continuous-improvement efforts in the plant depend on this information existing. *Estimate to be validated with the continuous-improvement lead*.

And the fair question from the production manager

“Is this a way to monitor my people?” — no: iLEAN listens to the shift lead, not to the operators, and only after the wake-word. Turning a dictated sentence into lathe, lot and symptom is an anchored task, where the best models drop below 1.5% error [1], and every structured issue can be reviewed and corrected by the lead before it is filed.

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