Back to all articles

Best predictive maintenance software for HVAC service contractors

How HVAC service contractors evaluate predictive maintenance software in 2026: sensor coverage, rules engines, alarm escalation, and EPA/DOE benchmarks.

KIContent TeamAug 27, 2026 — 9 min read
Best predictive maintenance software for HVAC service contractors

HVAC service contractors juggle chillers, RTUs, boilers, and VRF systems across dozens of client sites, and the difference between a scheduled truck roll and an emergency callback usually comes down to whether anyone was watching the equipment before it failed. This guide breaks down what predictive maintenance software for HVAC contractors actually does, what to check before you buy, and where condition-based monitoring fits next to your existing service software.

TL;DR
  • Predictive maintenance software for HVAC contractors works by tracking live sensor data against thresholds, not by guessing future failures with AI.
  • Look for sensor coverage across RTUs, chillers, boilers and VRF units plus a rules engine that maps to real HVAC failure modes.
  • The Kilo IoT Platform runs condition-based alarms on vibration, temperature and amperage data and escalates them to the right technician in 2026.
  • EPA Section 608 already requires leak-rate tracking on refrigerant systems over 50 lbs — software that surfaces that data saves a compliance headache.
  • A free tier covering 5 devices, 1 dashboard and 1 rule lets a contractor pilot condition monitoring on one site before buying in.
Numbers that shape the buying decision
20%
EPA Section 608 leak-rate trigger
Repair required within 30 days
~35%
Of commercial building energy use is HVAC
U.S. Department of Energy
€25/mo
Starter tier covers 25 monitored devices
Kilo Cloud pricing, 2026

What is predictive maintenance software for HVAC contractors?

Most software sold under the label "predictive maintenance" for HVAC is really condition-based monitoring: sensors report live data — temperature, vibration, current draw, refrigerant pressure — and a rules engine compares that data against thresholds you set. When a reading drifts, the system fires an alarm before the equipment trips a hard fault. That is a meaningfully different mechanism than machine-learning failure prediction, and contractors evaluating software should ask vendors directly which one they're getting, because the marketing language rarely draws the line.

The Kilo IoT Platform is built around the condition-based model: a visual rules engine using CEL expressions watches sensor streams and triggers alarms on thresholds you define, then routes them through an escalation chain to the tech on call. There's no black-box prediction happening — you set the rule, you see the logic, and you can step through it against a test payload before it goes live.

Predictive vs preventive vs condition-based HVAC maintenance: what's the difference?

Preventive maintenance is calendar-based — you service a chiller every 90 days whether it needs it or not. Condition-based maintenance replaces the calendar with live data: you service the chiller when vibration amplitude or discharge temperature actually drifts out of range. "Predictive" gets used loosely for both condition-based systems and true ML-driven failure forecasting, and the two require very different data volumes and modeling effort.

ANSI/ASHRAE/ACCA Standard 180 sets the baseline for documented HVAC inspection and maintenance practices in commercial buildings — it's the reference most facilities teams already work against for RTUs, boilers, and air handlers. Condition-based software doesn't replace that standard; it gives you the sensor data to justify skipping or expediting a task inside it.

What to look for in predictive maintenance software for HVAC service contractors

Sensor coverage across your actual equipment mix

A platform that only handles single-zone temperature sensors won't cover a service contract that includes chillers, boilers, and rooftop units. Check whether the vendor supports vibration, current/amperage, pressure, and refrigerant-adjacent sensors over LoRaWAN, mioty, or MQTT, not just a single proprietary sensor line.

A rules engine that maps to real failure modes

Generic "anomaly detection" dashboards are hard to trust because you can't see the logic behind an alarm. A visual rules engine with version control and one-click rollback lets a service manager write a rule like "compressor amp draw exceeds 15% of baseline for 10 minutes" and know exactly why an alarm fired.

Alarm escalation that reaches the right technician

An alarm that sits in an inbox nobody checks is worse than no alarm — it creates a false sense of coverage. Look for multi-step escalation chains, severity tiers, and quiet-hours logic so a low-priority drift doesn't page a tech at 2 a.m. while a compressor overcurrent event does.

API access to your existing dispatch or CMMS workflow

Most HVAC contractors already run a CMMS or field service platform for dispatch and billing. Condition-monitoring software doesn't need to replace that — it needs a REST API or webhook so an alarm can post a ticket into the system you already use, rather than forcing techs to check a second screen.

Per-device pricing you can scale with your book of business

A contractor monitoring 15 units at one client site has different economics than one running 300 units across a portfolio. Pricing tiers that scale by device count, with gateways excluded from the count, keep the math predictable as you add sites.

A digital twin or floor-plan view for multi-site accounts

When a service manager is covering ten buildings, a flat list of sensor IDs doesn't tell them which alarm matters. A dashboard with map or floor-plan pin widgets tied to a digital twin of each site turns a sensor ID into "chiller 2, mechanical room, Building C."

See condition-based HVAC monitoring live

Try the free tier on one site before rolling out across your book of business.

How condition-based monitoring actually triggers an HVAC service call

Here's the mechanical sequence, without the marketing gloss. A vibration sensor mounted on a chiller motor reports amplitude every few minutes over LoRaWAN. A CEL rule in the rules engine checks that value against a baseline threshold — say, a sustained 20% increase over the equipment's normal running range. When the threshold is crossed, the rule fires an alarm at the severity tier you configured, and the escalation chain notifies the on-call tech by SMS or push, with email as a backup channel if nobody acknowledges it inside a set window.

That's the model behind vibration monitoring for chiller plants: no failure forecast, just a documented, testable rule that catches drift before it becomes a hard trip. The same pattern applies to boiler room amperage, RTU discharge air temperature, or refrigerant-side pressure — the sensor changes, the rule logic doesn't.

The EPA's Section 608 refrigerant management rule requires that comfort-cooling appliances holding more than 50 pounds of refrigerant be repaired within 30 days once the annual leak rate hits 20%. A pressure or temperature threshold rule that flags early drift on a large rooftop system gives a contractor a documented trail toward that compliance requirement, rather than discovering the leak rate during an annual audit.

Predictive maintenance software for HVAC contractors: feature comparison

ApproachData sourceTrigger typeTypical lead time before failureBest fit
Calendar-based preventive maintenanceService log, not live dataFixed scheduleNone — reactive to schedule onlyLow-criticality equipment, tight budgets
Building automation system (BAS) trendingPoint trends inside the BASManual review of graphsDepends on how often someone checksSites with an existing BAS and dedicated staff
IoT condition-based monitoring platformWireless sensors (LoRaWAN, mioty, MQTT)Threshold or CEL rule, auto-escalatedHours to days, depending on the ruleMulti-site contractors covering equipment a BAS doesn't reach

Condition-based platforms like Kilo sit in the third row because the sensor layer is wireless and the equipment doesn't need to already be wired into a building automation system — a real advantage for retrofit work where running new BAS points isn't in the service contract.

Where to buy IoT sensors for HVAC predictive maintenance

Software is only half the deployment — you also need sensors rated for mechanical rooms, rooftop exposure, and cold equipment enclosures. Kilo Electronics, Kilo's sister hardware company, ships LoRaWAN and mioty sensors worldwide, including vibration and temperature units suited to chiller and boiler monitoring; check Kilo Electronics for current sensor availability before locking in a rollout plan.

Start with a single high-value asset — the chiller with the worst service history, or the RTU that's called in the most emergency tickets in 2026 — rather than instrumenting an entire portfolio on day one. A pilot at that scale fits inside the free tier: 5 devices, 1 dashboard, and 1 rule, no card required, and it doesn't expire. Once the rule proves out, the Starter tier at €25 per month covers up to 25 devices, with gateways unlimited on every tier so a site's radio coverage never counts against the device budget.

The refrigerant leak rule most HVAC contractors miss

Contractors focus service contracts on compressors and belts, but the EPA's 20% annual leak-rate threshold under Section 608 applies to comfort-cooling systems over 50 lbs of refrigerant regardless of how well the mechanical components are running. A pressure-drift rule that flags a slow leak weeks before an annual leak-rate audit turns a compliance surprise into a scheduled repair — and it's one of the simplest rules to write in a CEL-based engine.

FAQ

What is predictive maintenance software for HVAC contractors?

It's software that tracks live sensor data from HVAC equipment — vibration, temperature, current draw, pressure — against set thresholds and alerts a technician before a hard failure. Most platforms marketed as 'predictive' are actually condition-based, triggering on rules rather than forecasting failures with machine learning.

Is condition-based monitoring the same as predictive maintenance?

Not exactly — condition-based monitoring triggers alarms from live sensor thresholds, while true predictive maintenance implies statistical or machine-learning failure forecasting. Most HVAC software sold as 'predictive' in 2026 is condition-based, and contractors should ask vendors which mechanism they're actually getting.

How much does HVAC monitoring software cost?

Pricing varies by vendor and device count. Kilo's free tier covers up to 5 devices, 1 dashboard, and 1 rule with no card required, and the Starter tier is €25 per month for up to 25 devices, with gateways unlimited on every tier.

Can predictive maintenance software integrate with my CMMS?

Most condition-monitoring platforms don't replace a CMMS — they post alarm data into one via a REST API or webhook, so a threshold breach can generate a work order in the system your dispatchers already use.

What sensors do I need for HVAC condition monitoring?

Vibration sensors on motors and compressors, temperature sensors on discharge air and refrigerant lines, and current/amperage sensors on compressor circuits cover most failure modes. LoRaWAN and mioty sensors are common because they run for years on battery in mechanical rooms without new wiring.

Does the EPA require leak monitoring on HVAC refrigerant systems?

Yes — under Section 608 of the Clean Air Act, comfort-cooling appliances holding more than 50 pounds of refrigerant must be repaired within 30 days once the annual leak rate reaches 20%. Pressure or temperature drift rules can flag early leaks well before an annual audit.

Is a building automation system enough for predictive maintenance?

A BAS covers equipment that's already wired into it, but most HVAC contractors service assets outside that scope — standalone RTUs, remote boiler rooms, retrofit sites. Wireless IoT sensors extend condition monitoring to equipment a BAS never reaches.

How long does it take to set up HVAC predictive maintenance software?

A single-asset pilot — one chiller or RTU with a handful of sensors — can be live within a day once sensors are mounted and paired. Scaling to a full portfolio depends on how many sites need sensor placement and how many rules the service team wants to write.

You might also like