An alert that just says “value high” is noise — the plant already knows something is wrong. Munshi raises the deviation with the logged cause beside it, sends it to the person who owns that asset on WhatsApp or push, and escalates if it goes unacknowledged. The point is not the alarm; it is the right person, the moment it matters, with enough to act.
Your ERP reports the exception after the shift.
By then the operator who knew why is gone, and the moment to act has passed.
Why the gap existsWhat it does
A threshold or deviation fires with the operator’s logged reason beside it, so the alert is already halfway to an action instead of a prompt to go investigate.
Delivered on WhatsApp or push to whoever owns the asset — not buried in a dashboard nobody has open at 2 a.m.
An alert that goes unacknowledged climbs the chain on a timer. Nothing that mattered gets to sit unseen because the first person missed it.
In the plant
A furnace holding below setpoint raises to the shift in charge with the deviation and the logged reason attached — not a “value low” nobody reads.
A batch trending under DFT flags before the whole run needs a re-coat, to the person who owns the line.
An alert that goes unacknowledged climbs the chain on a timer. Nothing that mattered sits unseen because the first person missed it.
What Munshi surfaces
Teal is normal, maroon is flagged. Illustrative data.
Settling as thresholds are tuned — fewer, better alerts, not more noise.
By kind.
Alert by alert. The one over the line is where escalation took over.
Most alerts are handled by the first person. The rest is what escalation is for.
Get started
See this on your own floor.
Bring a month you already understand. The useful test is whether Munshi finds what you already know — and then what you didn’t.