Work
From observation to signed-off job, in one thread
A work order here is not a ticket that points at an asset. It is attached to the asset, carries the parts that came out of stock, the hours that were booked against it, the failure coding that makes it comparable, and a signature that cannot be edited afterwards.
A lifecycle the server enforces
Every transition is checked in the backend. Putting a job on hold or cancelling it requires a reason, not as a form hint, as a validation rule.
Raised, not yet started
Needs a decision before work begins
Someone is on it
Blocked. A reason is required
Signed off
Abandoned. A reason is required
Numbers are allocated atomically as WO-YYYY-NNNN and are never reused, so a reference in an email a year later still resolves to exactly one job.
What a job carries
ISO 14224 failure coding
Detection method, failure mode, mechanism, cause, four levels, with a vocabulary the server enforces. An out-of-vocabulary code is rejected, not quietly stored, which is the difference between reliability data and a free-text field nobody can aggregate.
Parts against real stock
Plan, reserve, consume, return, against the same inventory the storeroom sees, with serialised units followed individually from shelf to installed. Consumption moves stock; it does not just note that it happened.
5 signature roles, immutable
Technician, supervisor, safety, inspector, customer. One signature per person per role, stamped with a server-side UTC time, and not editable afterwards by anyone including us.
8 task types
Not just checkboxes: readings, numbers, pass/fail, range checks, multiple choice, photo capture. A completed inspection produces structured evidence rather than a tick and a memory.
Replace Bearing Assembly, Pump P-101
WO-2026-0042In ProgressBearing removed. Significant scoring on inner race, replacing with SKF 6310-2RS.
LOTO applied. Pump isolated and drained.
Beyond the single job
Preventive schedules on time or on meters
Generate on a calendar, on running hours, or on both. Seasonal months, advance-generation windows and a cap on how many open jobs a schedule may have, so a forgotten schedule cannot bury a crew.
Requests, separated from work
A simple intake form for whoever spots the problem, with its own numbering and a claim that is atomic, two people converting the same request cannot both win.
Reliability figures that come out of the work
Mean time between failures, mean time to repair, availability, planned-versus-reactive ratio, derived from jobs and failure codes rather than typed into a monthly report.
Jobs Powoflow raises itself
A vibration reading past a severity threshold, or an acoustic result flagged critical, raises a work order automatically with the priority mapped from severity and its origin recorded, so nobody has to notice first.
One number to be careful with
Preventive-maintenance compliance is reported as a percentage over a window you configure. Cancelled jobs stay in the denominator, which means a site with no preventive programme at all reports flatteringly. We would rather tell you that than have you build a board slide on it. Use it to spot a trend in your own numbers, not to compare two sites.
See it against your own failure codes
The fastest way to judge this is to bring a real job and a real part number and watch what the record looks like at the end.