NIGHTFOREMAN

Edge predictive maintenance for machine builders

Know what your machines are doing — years after they ship.

NightForeman is an industrial edge device that stands watch in your machine's control cabinet — reading what the PLC already measures and flagging developing faults before they become warranty claims. It never sleeps and never misses a shift. No cloud required, no inbound connections to your customer's network.

Built for OEMs with 50–500 machines in the field.

537,561
records collected
64
sensor channels
99.97%
data continuity
13
assets monitored
60 days
continuous operation
FIELD PILOT · EUROPEAN SLUDGE-DEWATERING FLEET · SCREW PRESSES, PUMPS, AGITATORS, DOSING · WINTER 2025–26

The problem

After handover, the machine goes dark.

You engineered it, built it, shipped it. From that day on, everything you learn about it arrives by phone — usually as a complaint.

WARRANTY

Claims you can't disprove

A machine fails after months of overload the operator never mentions. Without field data, you can't tell misuse from a factory defect — so you pay. A large share of warranty spend is unjustified, and unprovable.

REPUTATION

Downtime lands on your brand

When your machine stops a customer's process, nobody blames the sludge. Unplanned failures in the field cost you references, tenders, and repeat orders — quietly.

REVENUE

Service money left on the table

Your customers would pay for uptime. But without visibility into the fleet, you can only sell reactive repairs and scheduled checkups — not the monitoring contract they actually want.

How it works

Your PLC already knows. We stand watch.

One DIN-rail industrial PC in the cabinet. No new sensors to start, no changes to the machine, no inbound internet exposure to the OT network.

Install at the edge

The NightForeman box connects to the existing PLC/SCADA over standard industrial protocols and starts recording every parameter the machine already measures — torque, current, speed, pressure, flow.

Modbus TCPOPC UAoutbound-only

Learn on-site

Data is parsed and compressed 90–95% locally. An anomaly-detection model runs directly on the edge device and learns each machine's normal behaviour. Cloud sync is optional — never required.

edge ML≈10 s samplingno cloud dependency

Contextualize with operators

Flagged events appear on the machine's HMI, where operators label them in seconds: dense feed, blocked line, bearing noise. Every label teaches the model your machines' language — across the whole fleet.

HMI labelingfleet learning

For machine builders

Ship the machine. Keep the relationship.

NightForeman is white-labeled under your brand — your machines get smarter, and your service department gets a product to sell.

WARRANTY DEFENSE

Evidence, not arguments

A complete operating history of every unit: overloads, dry runs, out-of-spec feed. Settle misuse-vs-defect disputes with data, and target a 30–50% cut in unjustified warranty costs.

RECURRING REVENUE

Monitoring as a subscription

Turn one-time machine sales into a monitored fleet under contract. OEMs adding condition-based service typically grow service revenue 20–30%.

FLEET INTELLIGENCE

Field data back into R&D

See how your machines are actually operated across every site — which components run at their limits, which settings drift. Design the next generation on evidence.

SECURITY & COMPLIANCE

OT-safe by design

Outbound-only architecture: no inbound connections, no VPN into your customer's control network, aligned with NIST CSF and CISA guidance for OT environments. Selling into Europe? The system is already mapped to EU AI Act milestones.

Pilot program

Start with one machine.

A 60-day pilot on one production unit at a customer site: we install the edge device, build the baseline model, and hand you a report on what your machine has been trying to say. You keep the data.

Request a paid pilot TYPICAL INSTALL: ONE SITE VISIT · NO PRODUCTION STOP OR WRITE DIRECTLY: hello@nightforeman.com