Production Loss Analyst
Finds the ninety-second stops nobody writes down
- 8hard rules
- Things it must never do, enforced in every run
- 5handover rules
- Conditions where it stops and gives the work to a person
- 12output fields
- A structured result, not a paragraph you have to re-read
- ≤6steps per run
- Bounded work: at most 4 tool calls, enforced by the runtime
Read-only by design. This agent has no path to sending, spending, publishing or committing anything. Where that limit is the product, removing it would remove the reason to trust the output.
Overview
Manufacturers measure overall equipment effectiveness and then argue about the number instead of acting on it. It is the product of three things — availability, performance and quality — and reporting the product hides which one is costing you. A line at 62% could be stopping constantly, running steadily below rate, or producing scrap. Three problems, three owners.
The loss that dominates almost every line is the one nobody logs. Breakdowns get recorded because somebody has to be called. Short stops — a jam, a changeover overrun, a wait for material, ninety seconds at a time — are not worth writing down individually and are collectively larger than every breakdown combined. They appear in the data only as the gap between what the line should have produced while running and what it actually produced.
That gap is computable from records every plant already has, and it is where the money is. It is also why "we need a new machine" is usually the wrong conclusion: the machine is fine, it stops for ninety seconds forty times a shift, and nobody has added it up.
This agent computes that gap, splits every loss by cause, and prices each in units and hours so the largest one is obvious. It changes no setting and touches no machine.
Benefits
Capacity found without capital
Most lines have a shift's worth of output inside their own losses.
The right problem, with an owner
Maintenance, operations or quality — the split says which.
The machine you did not need to buy
Which is the usual conclusion when the gap is finally added up.
What it does
Computes the losses nobody logs
Short stops are individually too small to record and collectively larger than every breakdown combined. Visible only as a gap.
Splits the number instead of reporting it
Availability, performance and quality are three different problems with three different owners.
Prices each loss in units and hours
So the argument about the percentage ends and the largest loss is simply obvious.
Separates changeover from breakdown
One is a process you control and one is a failure. Merging them buries the fixable half.
Checks the rate itself
A line running at 100% of a rate set optimistically in 2019 is not running well.
Touches no machine
It reads records. It changes no setting, rate or schedule and has no route to any control system.
What this changes
What this changes about your week
- It takes over
- Working out where a shift's available time actually went.
- Done today by
- Argued about in a production meeting from a single percentage.
- How often
- Weekly, and the same argument each week.
- Typical manual effort
- The losses that dominate are never logged, so they have to be computed as a gap.
Why Manufacturing in particular. Short stops are individually too small to write down and collectively larger than every breakdown combined. This is why 'we need a new machine' is usually wrong — the machine stops for ninety seconds, forty times a shift.
Runs unattended
Started by you or by an event, and it finishes on its own. Nothing waits for someone to be at a desk.
The same standard every time
The two-hundredth item is held to the bar the first one was. Consistency is the part people cannot sustain.
It cannot act on its own
Production Loss Analyst has no path to sending, spending or committing. That limit is why its output is safe to act on.
How it runs
This agent runs server-side through the PROMIVO runtime. Each run is logged step by step and every tool call is permission-checked before it executes.
- 1InputValidated against the agent input schema
- 2Modelclaude-opus-5 with automatic fallback
- 3Toolsextract_structured_data
- 4PermissionsRuns READ actions automatically; anything higher waits for approval
- 5OutputStructured result matching the output schema
- 6LoggingTokens, duration and every step recorded
Demo
Demo dataIllustrative sample output, abridged.
Input
{
"line": "Line 2",
"runs": [],
"period": "Week 35"
}Output
{
"losses": [
{
"basis": "computed-from-gap",
"cause": "Short stops, individually unlogged",
"category": "performance",
"hoursLost": 41.2,
"unitsLost": 61800,
"shareOfLoss": 38.4
},
{
"basis": "logged",
"cause": "Changeover overrun",
"category": "availability",
"hoursLost": 18.6,
"unitsLost": 27900,
"shareOfLoss": 17.3
},
{
"basis": "logged",
"cause": "Breakdowns",
"category": "availability",
"hoursLost": 9.1,
"unitsLost": 13650,
"shareOfLoss": 8.5
}
],
"escalate": true,
"disclaimer": "An analysis of your own production records. No machine setting, rate or schedule has been changed, no control system was accessed, no safety determination is made, no individual is named, and losses marked computed-from-gap are derived rather than measured.",
"concentration": [
"62% of changeover loss comes from one product pairing, run 3 times a week. Sequencing it differently costs nothing.",
"Night shift short-stop loss is 2.3x day shift on the same product. Reported as a pattern with no attribution to any individual."
],
"overallPercent": 58.8,
"qualityPercent": 97.1,
"escalationReason": "Unlogged short stops are 38.4% of all loss and 4.5 times total breakdown time, and one product's standard rate appears understated.",
"performancePercent": 74.6,
"unloggedShortStops": "41.2 hours across the week, computed as the gap between standard rate during running time and actual output. That is 4.5 times all breakdowns combined, and none of it appears in any downtime log.",
"availabilityPercent": 81.2,
"unattributedPercent": 6.2,
"standardRateFindings": [
"Product B runs consistently at 104% of standard. The standard is likely understated, which flatters every performance figure that uses it."
]
}Requirements
Integrations
No integrations required.
Before you start
- Planned production time per shift or run
- Logged downtime with reasons
- Units produced and rejected
- Standard rate per product or line
- Requires the All-Access plan or higher
Use cases
Weekly production review
Where the shift actually went.
Before a capital request
Whether the capacity is missing or hidden.
A line that misses target
Which of the three losses is responsible.
Comparing shifts or lines
On comparable measures rather than on totals.
Pricing
All-Access
Billed monthly through your PROMIVO subscription. Cancel at any time.
Runs consume your plan allowance for agent executions and tokens. See plan limits
FAQ
Do we need sensors on the line?
No. It works from production records — planned time, downtime logged, units produced, rejects. The unlogged losses are computed as the gap, which is exactly why they are findable without new instrumentation.
Does it control any equipment?
No. It has no route to any control or automation system and changes no setting, rate or schedule.
What if our standard rates are wrong?
That is one of the findings. A line at 100% of a rate set optimistically years ago is not performing well, and it reports where the standard itself looks unreliable rather than trusting it.
Reviews
Reviews open once customers have run this agent.
Need this agent to work differently?
Tell us what to change and our team will scope a customised version for your business.
Customize this agent