Velosyti

Selling yearly service contracts with connected machines

The machine you already sold is quietly reporting run hours and faults. That data can become a yearly service contract, not just a warranty call.

· 3 min read · For machine builders and OEMs

Once a machine leaves your factory, you usually hear from the customer twice: at the warranty call, and when something breaks after the warranty ends. In between, the machine's own controller is quietly recording run hours, faults and motor load. Reading that data turns a one-time sale into a recurring service relationship, and gives you a genuine reason to call before something breaks, not after.

Signs this is worth doing

  • You already sell machines with a PLC or controller that logs run hours and faults, even if nobody reads that log today.
  • Your service calls are mostly reactive — a customer calls after a breakdown, not before.
  • Customers ask for an annual maintenance contract, but you price it as a flat guess rather than by actual usage.

What a good system does

  • Reads what the controller already knows — no new sensors needed on most modern machines.
  • Watches for a pattern, like motor load creeping up gradually, and raises a warning days before a failure, not after.
  • Gives the technician a specific likely cause, such as bearing wear, so they carry the right spare on the first visit.
  • Lets you price a service contract on actual run hours and condition, not a flat annual guess.

What it costs you

Most of the cost is a small connectivity box per machine and the setup to read your specific controller's data — cost that a service contract can recover within the first year for customers with more than a few machines. The larger shift is internal: your service team moves from reacting to breakdowns to reviewing warnings on a dashboard, which is a different daily habit than most teams have today, and is worth practising on a handful of machines before rolling it out across your whole installed base.

Questions to ask any vendor

  • Can it read our existing controllers, or does it need new hardware on every machine?
  • How early does a warning actually arrive before a real failure — days, or minutes?
  • Who owns the machine data — us, or the vendor?
  • Does it work with standard industrial protocols like OPC UA, or a closed system only they control?

How we can help

Thudippu, our machine health model, reads what a machine's controller already knows and warns before a breakdown, giving machine makers a real basis for yearly service contracts. See how this works for factories — AI for industry — or talk to us.

References

  1. Predictive maintenance — Wikipedia
  2. OPC Unified Architecture — Wikipedia