Automation and software built for the plant you actually have
Willowark builds the automation, inspection, and software layer between your machines and the people who have to make decisions about them. For most plants the constraint is not equipment. It is that the equipment does not talk, quality data lives on clipboards, and nobody can say what caused last Tuesday's bad shift. We remove that blind spot by measuring production where it happens and putting the result in front of the person who can act on it.
We work from the machine outward. Real signals first — PLC tags over EtherNet/IP or Modbus TCP, a dry contact off a stack light, current on a motor lead when a machine has no controller worth reading — normalized at the edge, then dashboards, scheduling logic, or inspection stations built on top. Controls, hardware, and software come from one team, so the handoff that usually kills these projects never happens.
A first project here is usually small and specific: monitoring on the two or three machines that gate everything else, or one inspection station where a recurring escape is costing real money. We install edge hardware, read whatever the machine will give us, and put a screen in front of the supervisor early in the engagement. That pilot settles what the equipment will actually report, how operators react to reason-code prompts, and whether the numbers change any decision. The plant-wide rollout is designed from what the pilot taught us, not from a proposal written before anyone opened a panel.
Reviewed
Sound familiar?
If you've said any of these, we should talk.
“We know we lost four hours yesterday. Nobody can tell me where.”
We put edge devices on the machines to timestamp every stop, then give operators a two-tap reason code on a panel or tablet. Downtime rolls up by machine, shift, part, and cause, so the Monday meeting argues about the top reason instead of whose number is right.
“Every time we quote a job we are guessing at the run rate.”
Automatic cycle capture turns actual run rate by part and machine into history instead of memory. Estimating pulls from the same table the floor produces, and the jobs that quietly lose money surface before you bid them again.
“Our best setup guy retires in eight months and it all leaves with him.”
We capture setup parameters, changeover steps, and troubleshooting decisions as structured records tied to each machine and part number. On top of that we can build an AI assistant that answers from your own SOPs, maintenance history, and drawings, citing the source.
“The MES quote came back at six figures and half of it we would never use.”
We build the part you need — usually job tracking, downtime, quality records, and a couple of integrations — and skip the rest. It runs on your hardware and you own it, with no per-seat licensing that grows every time you hire.
“Our IT guy does not want anything from the floor touching the office network.”
That is a reasonable position and we design around it. Edge devices sit on a segmented OT network with one-way data flow out through a gateway you control, no inbound access to controllers, and no cloud dependency unless you choose one. We document every port and protocol so IT can review it before anything is plugged in.
How this industry actually runs
The operation as we usually find it.
A working plant is rarely uniform. A 2003 press sits next to a 2021 robotic cell, one line runs on an SLC 500 nobody wants to touch, another on current ControlLogix. The ERP schedules against fixed cycle times while the floor runs to a whiteboard. Quality is sampled — first piece, then hourly — and traceability means a lot number on a traveler. Maintenance stays reactive because the only warning a machine gives is a noise one operator recognizes. Add turnover and customers who now want PPAP-grade documentation on parts that used to ship with a packing slip, and the shortage becomes information, not capacity.
Machine signals to the people who decide
Components:
- PLCs & sensors (counts, states, current)
- Legacy machine (dry contact / clamp)
- Edge gateway (normalize, buffer)
- Production dashboard (downtime, OEE)
- Alerts & reports (who acts, when)
Connections:
- PLCs & sensors to Edge gateway (EtherNet/IP, Modbus)
- Legacy machine to Edge gateway
- Edge gateway to Production dashboard (MQTT)
- Edge gateway to Alerts & reports
What we build
Starting projects that fit Manufacturing.
- Machine monitoring and OEE systems reading PLC tags, sensors, or bare stack-light signals
- Downtime capture with operator reason codes on a shop-floor HMI or tablet
- In-line vision for assembly verification, presence checks, and dimensional measurement
- Custom production software: job tracking, scheduling, quality records, and shipping paperwork
- ERP and accounting integrations so the floor and the office work from one set of numbers
- Edge gateways and historians that keep data local and keep logging when the network drops
- Maintenance alerting from vibration, current, and temperature sensors on the machines that give no warning today
- AI assistants that answer setup and troubleshooting questions from your own SOPs, drawings, and maintenance history
Capabilities we bring
Related work
Where this is running today.
Radar centering system for steel mills
A steel-mill systems provider · Steel manufacturing
Radar-based centering system for steel mills
End-to-end engineering of a radar sensing system that measures and centers material on steel mill lines — from equipment assessment through hardware selection, electrical engineering, software, installation, and commissioning.
Read the case study →X-ray thickness gauge
An industrial measurement company · Industrial measurement
X-ray thickness gauge
A complete X-ray based thickness gauge: equipment assessment, hardware selection, electrical and hardware engineering, API integration, all software, and commissioning — delivered as a working instrument.
Read the case study →Working in Manufacturing?
Tell us the line.
What runs by hand, what is not connected, what you are trying to build. An engineer replies within one business day with whether and how we would approach it.
Common questions
What Manufacturing teams ask first.
We only run a handful of machines. Is monitoring worth it?
Often more so, because one unexplained stoppage is a larger share of your output. A deployment on two or three constraint machines costs far less than a plant-wide rollout and usually pays back on quoting and scheduling accuracy alone.
Do you have to replace our PLCs to get data out?
Almost never. We read existing controllers over their native protocols, and where a machine has no useful controller we add non-invasive sensing such as a current transformer or a tap on the stack light. Your control program stays as it is.
How long before we see something working?
We scope a pilot on one line or cell that produces a usable result in weeks, not quarters. The pilot settles what specifications never do: what operators will tolerate, how noisy the data really is, and whether the metric changes any decision.
Who owns the system and the data once it is built?
You do. The software, source code, and the database it writes to are delivered to you and typically run on hardware in your building or in a cloud account you hold. There is no per-seat licensing and no arrangement where the data is only readable through our platform. We usually offer a support agreement afterward, but it is optional and the system keeps working without it.
Will operators actually use it, or will it become another screen nobody looks at?
That depends more on design than on the operators. Systems that get used usually ask for very little — a reason code in two taps, a confirmation at the end of a job — and give something back on the same screen, such as the count against target or the next job. We involve the lead operators during the pilot, because the ones who resist a system they were handed usually adopt one they helped shape.
Strategy. Software. Systems.
Engineering for Manufacturing.
Describe the problem in your own words. An engineer reads it — not a sales script — and tells you plainly what it would take.