Andon and Line-Stop Response, One Module of the ProAlert Platform

What andon is, plainly

Andon comes from the Toyota Production System. It is a signalling method that lets any worker stop the line and call for help the moment something is wrong, rather than passing a defect downstream and hoping it gets caught. Classically it is a light stack on the machine, an overhead board at the end of the aisle, and sometimes a pull-cord at every station.

The light was never the point. The point is that a problem becomes visible immediately and somebody qualified is obliged to come and deal with it. Everything worth digitising about andon follows from that: not the colour of the lamp, but who was called, who came, how long it took, and what they found.

What the digital version adds, and what it does not

A light stack tells the aisle that something is wrong. It cannot tell you who responded, whether three people walked over when one would do, how long the line waited, or whether the same failure has now happened eleven times this quarter. A digital call carries all of that as a record.

What it does not do is make the discipline work on its own. A plant with a poor escalation culture and a fast notification system is a plant that ignores notifications faster. The value is in the accountability the record creates, and in what the rest of the platform does with that record after the call closes.

How response coordination works in ProAlert

• Role-based routing

  A call is raised against an asset and routed to the roles configured for that asset... maintenance, quality, materials, supervision... through the Business Resource graph, so it follows the plant's actual reporting structure rather than a static distribution list that goes stale the week someone changes shift.
• Configurable call types per asset

  Each asset exposes only the call types that make sense for it, so an operator is not scrolling past nine irrelevant options to find the one that matches the problem in front of them.
• First-responder claim

  The first qualified responder to claim takes ownership and everyone else is released immediately. One person is dispatched instead of three. This is the single change most plants notice first, because duplicate walking is a cost nobody was measuring.
• Response and resolution timing

  Time to claim, time to arrive and time to resolve are recorded on every call. Response performance becomes a measured number per shift, per role and per asset instead of an argument between two people who each remember it differently.
• Responder route visualization

  While a call is open, the responder's position, distance to the asset and arrival estimate are visible, so a supervisor can escalate before a short stoppage turns into a lost shift rather than finding out afterwards.
• PIN-protected closure

  Closing a call requires the responder's personal PIN. The resolution is tied to a real individual, which is what turns a log into a chain of accountability that survives an audit.
• Structured failure classification on close

  Closure captures equipment, detail, cause and resolution against the ISO 14224 (Equipment Reliability Data Standard) taxonomy, and a failure analysis record is written on every close path... web, mobile and API. The record feeds failure mode analysis rather than dying in a free-text note.
• Physical buttons and existing boards

  A physical andon button wired to a GPIO input on a ProAlert EdgeSense device raises a call exactly as a tablet does, so existing pull-cords and buttons stay in service. Stack lights and boards can remain as the visual indication on the floor while the platform carries the record behind them.
• Emergency broadcast, separately

  Site-wide safety communication is a different mechanism from production calls: it reaches web, mobile and edge devices at once with acknowledgement tracking per recipient, and is backed by a safety incident register. It is not routed through the call queue, because a shelter instruction should never wait behind a material request.