Act on real faults sooner
Act on real faults sooner. The agent filters alarm noise and raises work orders with evidence attached.
Most BMS alarms are not faults. Repeats, nuisance alarms and sensors that flip in and out of alarm all day bury the one that matters, and by the time someone finds it, a valve has been wasting chilled water all morning. The alarm and fault triage agent watches every alarm across your sites, filters out the noise and turns real faults into work orders with the evidence attached.
These five short videos follow one day at Acme Facilities, from 48 alarms in ten minutes at Eastfield Mall to a chiller trip in the early hours that goes straight to a person. Facilio customers see 40% faster fault resolution with automatic work orders, and every critical fault and judgement call still reaches your team.
Watches alarms across every site
Alarms arrive from different BMS systems, chiller panels and meters across every site, each with its own screen and its own way of naming things. Someone has to watch them all, and across a large portfolio or overnight, that rarely happens. The person on shift has to work out which of dozens of alarms is real, usually by working down the list or waiting for a tenant to call. The alarms that matter sit in a list with hundreds that don't. By the time a real fault is spotted, it has often been running long enough for occupants to notice.
In this demo
0:00 It's 10:00 on a Tuesday morning at Acme Facilities. Alarms are coming in from building systems at fourteen sites: air handlers, chillers, pumps, meters and zone sensors, from three different BMS vendors.
0:16 The alarm and fault triage agent reads them all into one list. Each alarm is tagged with its site, its asset and its severity, using the same names your team uses in the CMMS, whichever system raised it.
0:34 At Eastfield Mall, things get busy. 48 alarms arrive in ten minutes: supply air temperature high on AHU-02, again and again, condenser pressure high on Chiller 1, a warm zone on level 2 and a sensor out of range.
0:55 For the person on shift, this is the hard part. Which of these 48 alarms is real? Most teams either work through the list from the top or wait for a tenant to call.
1:15 The agent starts sorting as the alarms arrive. It counts repeats, checks each alarm against your alert rules and looks at what the asset is doing right now.
1:30 Nothing is hidden: the full list stays on screen. In the next video, we'll see how 48 alarms become six issues.
Someone watches several BMS screens and works through the alarm list one alarm at a time, or waits for a tenant to call.
Every alarm from every site lands in one list, and the agent sorts it as alarms arrive.
Filters repeats, nuisance and flapping alarms
The same alarm firing every minute, a chiller alarm that clears on its own, a sensor that flips in and out of alarm all day: these make up most of an alarm list. Each one still has to be looked at, acknowledged or dismissed by someone on shift, and the list never gets shorter. Engineers learn to ignore them, because nearly every time they check, there's nothing to fix. That habit is exactly how a real fault gets missed. Turning the noisy alarms off isn't the answer either, because the next time one of them is real, nobody will hear it.
In this demo
0:00 Back at Eastfield Mall, 48 alarms have arrived in ten minutes. Let's watch what the agent does with each one.
0:22 First, the repeats. Supply air temperature high on AHU-02 has fired twelve times. The agent folds them into one issue with a count of twelve. Across the list, 42 repeats fold away, and 48 alarms become six issues.
0:58 Next, condenser pressure high on Chiller 1. The reading came back to normal four minutes later and has stayed there. The agent closes it automatically and records why: no work order needed.
1:18 Then two alarms inside your alert rules: a zone on level 2 at 24.6 degrees, and a small filter pressure drop on AHU-01. Neither crosses the line you set for raising work, so the agent holds them and keeps watching.
1:40 Sensor T-114 has flipped between normal and out of range fourteen times this morning. That's a flapping sensor, not a building problem. The agent holds the alarms and marks it as a likely sensor fault.
2:00 Every one of these decisions sits in the triage log with its reason. Nothing is deleted, and anyone on the team can see what was held and why.
Engineers scroll past repeats and nuisance alarms all day, and a real fault gets lost among them.
Repeats are folded, alarms that clear are closed and nuisance alarms are held, each with a logged reason.
Groups related alarms to one likely fault
One fault rarely raises one alarm. A stuck valve can show up as temperature alarms, valve alarms and an energy spike, each listed separately and sometimes each with its own work order. Engineers chase the symptoms one by one while the cause keeps running, and the building keeps paying for it. Working out what links them means pulling up schedules, live readings and trends for the asset, which takes time few teams have in the middle of a busy shift. So the same fault can produce several visits before anyone finds what's actually wrong.
In this demo
0:00 Two of the six issues at Eastfield Mall are on the same air handler, AHU-02: the cooling coil valve reported open four times, and supply air temperature high twelve times. The agent checks whether they share a cause before anyone is sent.
0:24 It starts with the asset's schedule. AHU-02 serves the food court, which is closed for a deep clean this morning, so the supply fan was stopped on schedule at 6:40.
0:45 Now the live readings. The supply fan VFD shows zero speed, as planned. But the chilled water valve is still at 100%, and it should close when the fan stops.
1:08 The trend over the last 24 hours confirms it. On other mornings the valve closed within a minute of the fan stopping. Today it opened fully and stayed there.
1:30 That points to one fault: the coil valve is open while the fan is off. The supply air alarms are a symptom. With no air moving, the sensor is reading the warm plant room.
1:55 The agent links all 16 alarms to that one fault, so the team sees one problem to fix instead of two issues and sixteen alarms.
2:15 It also estimates what the fault costs while it runs: about 38 kilowatt hours a day of chilled water energy, for a fan that isn't moving any air. That goes into the evidence.
2:30 The diagnosis is the likely cause, with the readings to back it. The engineer who attends confirms it on site.
Each alarm is chased on its own, and the same fault can produce several work orders before anyone finds the cause.
Related alarms are linked to one likely fault, checked against live readings and trends, with the reasoning shown.
Raises the work order with evidence
When a real fault is found, someone still has to write it up: which asset, what's wrong, how urgent it is and what the technician should look at first. Most alarm work orders arrive with a one-line title copied from the alarm, and the technician starts the diagnosis again on site. That can mean a wasted first visit, a second trip with the right parts and a fault that runs longer than it needed to. The evidence that would have saved the trip usually existed in the BMS trends, but nobody had time to attach it.
In this demo
0:00 The agent has found one real fault at Eastfield Mall: the cooling coil valve on AHU-02 is stuck open. Now it raises the work order, straight into the CMMS.
0:14 Work order #72: AHU-02 cooling coil valve stuck open. Priority is high, set by your rules for faults that waste energy on a key asset.
0:30 Here's the evidence. The likely root cause: valve open while the fan is off. The impact: about 38 kilowatt hours a day. And the 16 alarms linked to it, including the supply air temperature alarms.
0:52 The trend charts are attached too: valve position against fan speed over 24 hours, so the technician sees at a glance what the agent saw.
1:08 The agent also notes what it ruled out. The fan stopped on schedule, so this isn't a fan fault, and the temperature sensor is reading as expected for still air.
1:20 The job goes to Acme's HVAC team, the team your rules assign to air handling at Eastfield. Sam picks it up and heads to the plant room knowing what to look for.
1:40 When Sam closes the job, the readings, the linked alarms and the fix sit on one record. One fault, one work order, one visit.
Alarm work orders arrive with a one-line title, and the technician starts the diagnosis again on site.
Each work order arrives with the likely cause, impact, linked alarms and trends attached, routed to the right team.
Escalates critical faults to a person
Some faults can't wait for the morning queue: a chiller trip on a server room, a cold store warming up, a pump failure at the weekend. These need a person's judgement now, not a work order waiting to be picked up. Overnight and at weekends, a critical alarm can sit in a list until someone happens to see it or a tenant calls. Even when the right person is reached, they often start with nothing but an alarm name and have to log in to find out what's happening. Every minute spent working that out is a minute the problem gets worse.
In this demo
0:00 It's 2:10 on a Sunday morning. At Quayside House, Chiller 1 trips. It cools a tenant's server room, and your escalation rules mark that as critical.
0:20 The agent doesn't fold this into the queue. It checks the live readings first: chilled water supply has climbed from 7 to 11 degrees in eight minutes, and Chiller 2 is on standby.
0:40 It alerts Tom, the on-call engineer, straight away. The alert says what tripped, what it serves, the readings so far and the trend, so Tom knows what he's dealing with before he opens his laptop.
1:00 At the same time it raises a work order marked critical with the evidence attached, so the record starts now, not on Monday morning.
1:35 Tom hasn't acknowledged within the ten minutes your rules allow. The agent moves to the next name on the list and alerts Leena, the duty manager, with the same evidence and a note that Tom hasn't responded.
1:58 Leena acknowledges and opens the evidence. Chiller 2 is still on standby, and the server room is warming but still inside its limit.
2:20 The decision is hers. She asks site security to start Chiller 2 and books an engineer for first light. The agent logs each step against the work order.
2:45 Critical faults go to a person, every time. The agent's job is to get them there quickly, with the evidence in hand.
A critical alarm waits in the list overnight until someone happens to see it or a tenant calls.
Critical faults go to the on-call engineer straight away with the evidence, and up the list if nobody responds.
How the alarm and fault triage agent goes live with your team
Our FM Ops Engineers build the alarm and fault triage agent around your alarm rules, sites and escalation lists, and run it with you from day one. Your team needs no training, and real faults reach the right people sooner with the evidence attached.
- 01 We learn your operation Our FM Ops Engineers sit with your team and map how work is assigned today.
- 02 We build the agent Your rules, sites and vendors, built into the agent with you.
- 03 We run it Deployed, maintained and fine-tuned by our team. Your team gets the outcome.
