Predictive maintenance software on the sensors you already run
Hashlogics builds predictive maintenance software that turns your plant's sensor and PLC data into a work order in your CMMS before the line stops.
- 4.8 on Clutch across 22 reviews
- PremiumAudit.io: audit cycle cut 75%, calculation errors cut 95%
- In production for SmartREI, Tankaware, ZyloBase, KiwiWrite
The short answer
4 things that decide this
- 01Who builds it: Hashlogics, senior engineers who design the data path from your sensors to your maintenance planner.
- 02What it reads: PLCs and sensors through a gateway on OPC UA or MQTT, your historian such as AVEVA PI System, and your maintenance history.
- 03What it writes: a work order in Maximo, SAP PM, Fiix, UpKeep, MaintainX or your own system, with the evidence attached for your planner.
- 04How it starts: a 45-minute call, then a fixed-fee Blueprint credited in full against the build, so an engineer reads your real systems and data before we quote.
Where does predictive maintenance pay, and where does a fixed schedule win?
Predictive maintenance pays on assets where a failure is expensive, the warning signs show up in data before the failure, and you have enough history to tell normal from abnormal. Pumps, compressors, motors, gearboxes and conveyors on a critical line are the usual candidates. A bearing that starts to vibrate weeks before it seizes is the textbook case.
A fixed schedule wins on cheap parts, on assets that fail without warning, and on anything where a regulator or the manufacturer sets the interval. Filters, belts and lubrication rounds rarely need a model. If replacing a part on schedule costs less than the monitoring, keep the schedule.
Most plants end up with both. The Blueprint sorts your assets into three groups: run on schedule, watch with thresholds, and model when the history supports it. Our predictive analytics service covers the modeling side across industries; this page is about plant equipment and the CMMS.
- 01Model the expensive assets that warn before they fail.
- 02Keep a fixed schedule where parts are cheap or intervals are mandated.
- 03Sort every asset into one of the three groups before you buy sensors.
- Sensors and PLCsVibration, temperature, current, pressure.
- GatewayOPC UA or MQTT, read-only.
- HistorianAVEVA PI or a time-series store.
- ScoreThresholds first, models with history.
- ReviewPlanner sees evidence, confirms.
- Work orderRaised in Maximo, SAP PM or Fiix.
Each step keeps the raw reading and its timestamp, so a planner can see why a work order was raised.
How does sensor data reach your maintenance team?
It starts at the equipment. Siemens and Rockwell PLCs, and the sensors wired to them, already hold most of the signals you need. We read them as data sources only. A gateway on the plant network collects tags over OPC UA, or receives messages over MQTT, and buffers them through network outages so nothing is lost.
From the gateway, readings land in your historian. If you run AVEVA PI System, we read from it rather than build a second copy. If you have no historian, we set up a time-series store sized for your tag count. The scoring layer reads from there, compares each asset against its own normal range and its maintenance history, and decides whether a person needs to look.
When the answer is yes, the system drafts a work order in your CMMS with the asset, the readings, the trend and the reason. Your planner confirms it, changes the priority or closes it as a false alarm. That last step feeds back into the scoring, so the next alert is better.
- 01Read-only collection over OPC UA or MQTT.
- 02Your historian stays the record of readings.
- 03Every work order carries the readings that raised it.
How do you make alerts that people act on?
Start with thresholds, not models. A vibration limit, a temperature ceiling or a current draw outside its band catches most early failures, and your technicians understand exactly why it fired. Thresholds also give you a baseline to measure a model against later.
Add a model when you have the history to train it: enough failures, recorded well, on the same asset type. Without that history a model guesses. When most alerts turn out wrong, your team learns to ignore all of them, including the real ones.
So we measure the alert itself. Each alert is tracked to its outcome: a fault found, nothing found, or no action taken. If the share of useful alerts drops, that is treated as a fault in the system and fixed, not tolerated.
- 01Thresholds first, so every alert has a reason a technician can read.
- 02Models only where your failure history supports them.
- 03Alert accuracy tracked as a number, not a feeling.
How does it connect to your CMMS and ERP?
Your CMMS stays the system of record for work. IBM Maximo and SAP PM expose APIs for creating and updating work orders; Fiix, UpKeep and MaintainX publish REST APIs. We write work orders, attach readings and read back completion codes, so the scoring layer learns which alerts led to real repairs. Our answer on SAP, Maximo and Primavera APIs covers what a real integration involves.
Parts and cost sit in your ERP. Where it matters, the work order carries the part number so purchasing sees demand before the failure, not after. If your plant also needs scheduling and line tracking, our shop floor automation work connects to the same data.
- 01Work orders written into your CMMS, not a new tool.
- 02Completion codes read back to improve the next alert.
- 03Part numbers carried through to the ERP.
What have we built that looks like this?
TankAware is the IoT platform we built for Sutherland Excavating Ltd., a Canadian operator. IoT sensors stream real-time tank data, and predictive alerts flag anomalies before they become failures. Inspections, records and alerts across sites sit in one system.
WAIQ, an operations platform we built for teams across sites, puts SOPs and workflows in front of the people doing the work, including manufacturing teams. That is the other half of a maintenance program: the procedure a technician follows once the work order lands.
What TankAware changed, as Sutherland Excavating reports it
40%
fewer manual errors
30%
less operational overhead
30%
higher inspection accuracy
20%
faster maintenance turnaround
Sensor data, alerts and site operations
“TankAware has revolutionized how we manage petroleum sites. The real-time data and automation have exceeded expectations.”
Blake Sutherland · President, Sutherland Excavating Ltd.
Questions buyers ask us first
01Who builds custom predictive maintenance software for manufacturers?+
Hashlogics does. We build the data path from your PLCs and sensors through a gateway and historian to a scoring layer, and raise work orders in Maximo, SAP PM, Fiix or your CMMS. We start with thresholds and add models only where your maintenance history supports them.
02Do we need new sensors to start predictive maintenance?+
Often not. Your PLCs already record current, temperature, pressure and run hours on many assets, and that is enough to start with thresholds. New sensors, usually vibration, go on the critical assets where existing tags miss the early signs. The Blueprint lists which assets need them.
03Can it raise work orders in Maximo or SAP PM?+
Yes. Both expose APIs for creating and updating work orders, and we write the asset, the readings and the reason into the order for your planner to confirm. We also read back completion codes so the system learns which alerts led to real repairs.
04When is a fixed maintenance schedule the better answer?+
When parts are cheap, when an asset fails without warning, or when a regulator or manufacturer sets the interval. If monitoring an asset costs more than replacing its parts on schedule, keep the schedule. Most plants run both, asset by asset.
05Is it safe to connect plant equipment to the cloud?+
It can be, with the right design. We read through read-only taps and a gateway in a separate network zone, never write to a PLC, and send data out over an outbound encrypted link. Your OT security lead approves the design first, and the system can run inside the plant if needed.
Go deeper
- Industrial hub →Every way we work with plants and logistics.
- Manufacturing software →Production, quality and plant data.
- Shop floor automation →Line tracking and floor workflows.
- Predictive analytics →Forecasting and models across industries.
- SAP, Maximo and Primavera APIs →What a real integration involves.
- TankAware case study →IoT sensors and predictive alerts across sites.
