How to Automate UpKeep

A work order is the only record that a machine was ever touched

UpKeep is a CMMS: where a maintenance team keeps the history of every machine, pump, chiller, forklift and building it is responsible for. Manufacturing plants run on it, and so do facilities teams, utilities, food producers, schools, gyms and property managers.

The unit of work is the work order. Somebody reports a leak, a scheduled service comes due, a line goes down at two in the morning, and each becomes a job with an asset attached, a technician assigned, parts drawn from a stockroom and a note of how long it took. Technicians work it from a phone on the floor, offline if the plant has no signal down that aisle.

Underneath sits the asset register: every piece of equipment with its serial number, its location, its service history and the running total of what it has cost to keep alive. That register is why a plant can say which of two identical compressors to replace. A maintenance department without one is guessing in public.

Maintenance runs on parts and people the CMMS cannot see

The work orders are the visible half of the job. The other half happens on a screen that is not UpKeep.

A technician needs a seal and the stockroom shelf is empty. Somebody opens the distributor's site to see whether it is in stock and what the lead time is, then a second distributor because the first quoted six weeks. A pump is under warranty, possibly, and finding out means the manufacturer's registration lookup and a serial number typed in by hand.

Then the outside contractors. The lift company, the fire suppression inspector, the HVAC vendor on a service agreement. Each has a portal, or an email thread, and each was supposed to be on site this quarter. Statutory inspections carry their own dates and their own authority, and a boiler certificate that lapsed is not a maintenance problem until it is a shutdown.

None of this is skilled work. All of it is somebody senior at a desk with eleven tabs open, while the radio asks where the seal got to.

UpKeep

Lead times before the shelf is empty Every part below its minimum checked against your distributors' own stock pages, so an eleven week wait becomes a purchase decision this week.

Repairs that were still under warranty Serial numbers run against the manufacturer's registration lookup before work is authorised, rather than after the invoice has been paid.

Contractors held to the contract Lift, sprinkler and chiller visits confirmed on each vendor's own portal, so a service that was billed and never happened becomes an issue.

A PM trigger fires on a meter UpKeep is already reading

Inside its own walls UpKeep automates well, and a team still raising preventive jobs by hand is making work for itself.

Preventive maintenance triggers three ways: time-based on a calendar interval, meter-based on hours, miles or cycles, and condition-based on live sensor readings. When a trigger fires, a work order is created and assigned with the procedures, parts and priority attached, so nobody rebuilds it from scratch. Parts and inventory management tracks what is on the shelf.

Read that trigger list again and the boundary is obvious. Time, meters and sensors are all things UpKeep either holds or is wired into.

It knows the shelf is down to two seals. It has no way of knowing that the distributor now quotes eleven weeks on that part, that the warranty on the pump expires in March, or that the fire inspector last came fourteen months ago. Those facts live on other companies' websites, and none of them will ever send UpKeep a message.

Spare parts that report their own lead times

A reorder point can be a number about the world outside the storeroom, not just about the shelf inside it.

Every part below its minimum checked against your distributors' own stock pages, so a seal that has quietly moved to an eleven week lead time is a purchasing decision this week rather than a line down in October. Serial numbers run against the manufacturer's warranty lookup before a repair is authorised, because paying for work that was covered is the most avoidable money a plant spends.

Service contractors checked against what they were paid for: the lift inspection, the sprinkler test, the chiller service, each confirmed as done on the vendor's own portal and raised as an issue when it was not. Statutory inspection and certificate status read where the authority publishes it, so an expiry surfaces in weeks rather than during an audit.

Manufacturer service bulletins watched for the models actually in your asset register, so a known failure mode arrives as a scheduled job instead of as a breakdown.

Technicians were hired to fix things, not to check portals

Somebody in the maintenance office already holds all those logins. They get opened when there is time, which is why a lead time is discovered the week after the part was needed.

WebRun is an agent that works a real Chrome browser, signed in as you. It opens the distributor's stock page, the warranty lookup, the contractor's portal or the inspection authority's register, reads what is on the screen, and writes the answer back into UpKeep as a request, a work order or a note against the asset.

No API has to be arranged first, which matters because most parts distributors will never publish one. It runs on your schedule, you can watch a run and stop it, and anything that orders stock or books a vendor waits for you.

The workflows below are already built, and each one names the sites it opens.