Manufacturing software built for the plant floor
Three production systems built for multi-site industrial operations. Sensor data that means something, one method everywhere, and nobody working a shift they should not.
A plant runs on habits nobody wrote down. The second site does it slightly differently, the third differently again, and by the fourth nobody can say which way is correct. WAIQ exists because that drift costs more than any single machine failure.
What industrial engineering actually demands
4 things that decide this
- 01WAIQ holds SOPs, workflows and training across every site, department and role. Its own users span manufacturing, healthcare, retail and logistics.
- 02TankAware streams sensor data with early alerts across hundreds of industrial sites, tuned per site so the alarm keeps meaning something.
- 03Shift Link decides who can work before the rota goes out, using document checks and expiry dates rather than a report read later.
- 04All three were built multi-tenant from day one. Adding site isolation to a single-tenant system later is a rebuild.
Systems running unattended across sites
Three builds, three industrial problems
Machine data, method drift and shift eligibility are the three that quietly cost a multi-site operation money.
| Build | What it does | The plant problem it solves | Where it runs |
|---|---|---|---|
| TankAware | IoT telemetry with predictive alerts. | Knowing a machine is drifting before it stops. | Industrial sites, Canada |
| WAIQ | SOPs, workflows and training with readiness tracking. | Every plant working to the same documented method. | Multi-site operations, multi-region |
| Shift Link | Document validation gating shift eligibility. | Nobody on the line without current certification. | Workforce compliance, United Kingdom |
- SenseReadings arrive, some wrong.
- BaselineNormal differs per line.
- AlertTo someone who can act.
- MethodOne documented way to fix it.
- QualifyOnly cleared people do the work.
- ProveAudit trail, per site.
Most industrial software buys the first three and assumes the last three. The drift lives in the last three.
An alarm that cries wolf is worse than no alarm
Readings move for reasons that have nothing to do with a fault. Ambient temperature, a shift change, a batch running hot. Set your threshold against a fleet average and you will alert on the weather.
Model each line's own normal first, then alert on deviation from that. We stayed on TankAware's predictive models as sites came online, because a model tuned across ten locations is not tuned for the eleventh.
- Baseline per line, per season. A global threshold guarantees noise somewhere.
- Separate a drifting sensor from a drifting machine. They need different people.
- Track whether each alert was acted on. One nobody answers is a defect to fix.
The layer above your machines
Your machines and their controllers already do their job. They will not tell you whether the fourth plant works the same way as the first. Nor whether tonight's operator is still certified.
That is the layer we build, and it connects to what you already run. Sensor data is read from the equipment on your sites rather than replaced, and the operating record sits above it.
- 01Telemetry with alerting tuned per line, and an alert that becomes assigned work.
- 02SOPs, workflows and training with readiness visible per site, department and role.
- 03Eligibility gating so certification expiry changes who can be scheduled.
- 04Multi-tenant isolation, role-based access and audit trails your assessor can read.
The stack behind our industrial builds
Data and models
Platform
Controls
“TankAware has revolutionized how we manage petroleum sites. The real-time data and automation have exceeded expectations.”
Blake Sutherland · President, Sutherland Excavating Ltd.
What plant leaders ask us
01Our line data already sits in a historian. Why add software?
Because a historian records and does not act. It cannot turn a deviation into an assigned job. It cannot hold one method across four plants, or stop an uncertified operator being scheduled. That is the layer above it.
02Can a tablet on the floor work when the network drops?
Yes, if offline is designed in from the start. The entry is written to the device with its own timestamp and queued, then reconciled on reconnect. Both clocks are kept, so nobody can argue the record was fabricated later.
03How do you stop sensor drift corrupting our numbers?
By treating a drifting sensor as its own event class rather than a reading. Drift is more dangerous than outright failure, because a device two percent low for months quietly poisons the baseline everything else is judged against.
04Can we make every plant follow the same process?
WAIQ is the build that does exactly this. One set of SOPs, workflows and training runs across every site, department and role, with readiness visible per person. Allow local variation inside a fixed frame, because forcing four plants onto one identical script is how adoption dies.
05How do we know only qualified people are on the line?
Make certification expiry drive the schedule. Shift Link validates documents, tracks expiry, and removes workers from eligible lists automatically. A dashboard that flags a lapsed ticket after the shift has already failed at the only job that mattered.
