Solutions / Maintenance work order automation
Maintenance work order automation, written into your CMMS
For maintenance teams in mining, heavy equipment fleets and manufacturing whose work orders still start as a phone call, a line on a pre-start sheet or an alarm someone has to notice. Mirage agents read those signals, draft the work order with the asset, the symptom, its history and a suggested priority, send it to the planner for approval and follow it to closure in your CMMS or ERP maintenance module.
At Transwin, Mirage built per-machine tracking of greasing, oil changes and mileage on the mining equipment.
The reality
The sign that a machine needs work is usually already recorded somewhere: engine hours past a service interval, a telematics fault, an operator's note that the boom lifts slowly, a failed item on an inspection sheet. Turning it into a work order means someone noticing it, finding the asset in the CMMS, looking up what was done last time, choosing a priority and typing it in. When nobody does, the first work order is written after the breakdown.
What Mirage does
From a signal to an approved work order
Mirage does not replace your CMMS. It sits in front of it, reads what the operation already produces and writes the work order where your planners already work.
Signals read where they already are
Machine hours and usage data, telematics events, operator reports and pre-start sheets, inspection notes, even a message in a chat group. Each one is tied to an asset in your register, and reports of the same symptom on the same machine are grouped into one request.
A work order drafted, not a ticket
The asset and its location, the symptom in the operator's words and in your failure codes, the last work orders on the same asset and the parts used then, and a suggested priority with the reason for it. The fields, types and statuses are yours, mapped once.
Planner approval, then tracking to closure
The planner approves the draft, changes its priority or rejects it with a reason, and only an approved order is written into the CMMS. The agent then follows it: waiting for parts, past its planned date, closed without a completion note. Every action is logged with its inputs, sources and result.
What to settle with your planners first
The agent applies your maintenance rules; it does not invent them. Four things are agreed before the pilot.
Which signals open a work order
A service interval reached, a fault code, a reported symptom, a failed inspection item. Each rule names the signal, the threshold and the asset class it applies to. A signal no rule covers goes to a person as a question, not as an order.
How priority is set
On your criteria: safety, production impact, whether a standby machine exists, the time to the next planned stop. The agent suggests a priority and shows which criterion set it; the planner keeps the decision.
Where the order is written
Your CMMS or the maintenance module of your ERP, through its API where it has one and through a connector where it does not. Mirage names no CMMS vendor as a published integration; the connection is scoped on your system at the first meeting.
When an order counts as closed
Closed with a completion note and the hours or parts recorded. An order closed without them, or still open past its planned date, goes back to the planner instead of disappearing from view.
WHERE IT SITS
CMMS, predictive maintenance, work order agent: which does what?
This page covers one step, from a signal to an approved and tracked work order. The table places it among the systems around it.
| Layer | What it does | How it relates to this page |
|---|---|---|
| Your CMMS or ERP maintenance module | The system of record: assets, work orders, maintenance history, schedules | Stays where it is. The agent reads its history and writes approved orders into it |
| Predictive maintenance | Estimates when a component is likely to fail, from usage or sensor data | One possible signal among others; the blog's predictive maintenance guides cover it |
| Work order agent | Turns a signal into a drafted work order, sends it for approval, tracks it to closure | This page |
| Operations control room | A live view of the operation, where an open work order is one event among others | The control room page; mining operations software covers the per-shift record of a mine |
A prediction model is not a precondition. Most of the signals that open work orders, service intervals and operator reports, already exist without one.
The output
A maintenance work order, field by field
An illustrative work order, with invented values: what the agent read, what it drafted for the CMMS, and where each field stands.
| Field | From the signal | In the CMMS work order | Status |
|---|---|---|---|
| Asset | Telematics unit on loader L-07 | Asset LDR-007, wheel loader, pit 2 | Matched to the register |
| Symptom↳ The temperature alarm and the operator's note described the same problem on the same machine. They became one request, not two. | Hydraulic oil temperature alarm, 3 times this shift; operator note: boom slow to lift | Hydraulic system, overheating, boom slow to lift | Grouped from 2 signals |
| History | Not in the signal | Hydraulic filter replaced 6 weeks ago; same symptom reported once since | Read from the CMMS |
| Priority↳ The agent suggests the priority and names the reason: no standby loader in that pit. The planner decides, and the order is written into the CMMS only once approved. | Loader working, no standby loader in pit 2 | Suggested: high, before the next shift | Awaiting the planner |
| Release | Not in the signal | Approved by the planner at 06:10, written to the CMMS | Tracked until closed |
Proof
Equipment maintenance in production at Transwin
Mirage has not published a deployment that drafts work orders in a CMMS. The nearest published deployment is at Transwin, a Moroccan logistics operator with mining equipment: Mirage built per-machine tracking of greasing, oil changes and mileage, then layered prediction on top, so maintenance is scheduled before breakdown. Transwin's case study publishes no maintenance figure, and this page quotes none.
Transwin (mining equipment)
Per-machine tracking of greasing, oil changes and mileage
Prediction layered on top: maintenance scheduled before breakdown
5 systems in production, across 550 trucks and 100+ machines
Work order automation, on your own maintenance data
Send us a month of work orders exported from your CMMS, with the operator reports, inspection sheets or telematics exports behind them if you have them. We come back with the work orders we would have drafted from those signals, field by field, the priority we would have suggested and why, and what writing into your CMMS involves.
How a deployment runs: Observe, Map, Build, Run, Expand. See the method
Prefer to talk it through first? Book a call
FAQ
Frequently asked questions
Turning the signals that a machine needs work, usage hours, telematics events, operator reports and inspection notes, into work orders without someone retyping them. Mirage's agents draft each order with the asset, the symptom, its history and a suggested priority, send it to the planner for approval and track it to closure in your CMMS.
No. The CMMS stays the system of record for assets, work orders and history. Mirage writes into it, or into the maintenance module of your ERP, through its API or through a connector when it has none, using your work order types, statuses and fields.
Not on its own. Predictive maintenance estimates when a component is likely to fail. This page is about what happens next: any signal, a prediction, a service interval or an operator's report, becoming an approved and tracked work order. A prediction model can be one of the signals, but it is not required.
Your planner. The agent suggests a priority and shows the criterion behind it; the planner approves, changes or rejects the draft, and only an approved order is written into the CMMS. Every agent action is logged with its inputs, sources and result.
Mirage has not published a deployment that drafts work orders in a CMMS. Its published maintenance deployment is at Transwin: per-machine tracking of greasing, oil changes and mileage on the mining equipment, with prediction layered on top so maintenance is scheduled before breakdown. No maintenance figure is published.
Mining and quarry equipment, heavy equipment and truck fleets, and manufacturing lines: wherever machines produce usage data or operators report problems, and a CMMS or an ERP maintenance module holds the work orders.
A proof of concept on your own maintenance data one week after the first meeting, and the first workflow in production in 4 to 6 weeks. Work spanning several sites or systems can take longer and is scoped at the first meeting.
A fixed-price pilot on one asset class or one site first, which you can stop after. Then an annual subscription priced by volume, quoted in the first conversation.