What facility intelligence actually means

Most facilities already have data. Utility bills, a building automation system, maybe a handful of submeters and a spreadsheet someone updates monthly. What they do not have is a single operating picture that tells them what is happening right now and what to do about it. The gap between having data and being able to act on it is where most facility programs stall, and it is the gap the phrase "facility intelligence" is meant to describe.

The term gets used loosely, mostly by people selling dashboards. So it is worth being precise about what separates a reporting tool from an operating system for a building.

Reporting looks backward. Intelligence operates on the present

A monthly energy report tells you what you spent. It arrives after every decision that produced the number has already been made. You cannot unspend a demand peak that happened three weeks ago, and you cannot go back and catch the chiller that ran all weekend because someone left it in hand.

Facility intelligence collapses that delay. The measurement, the analysis and the notification happen close enough to the event that a human can still change the outcome. That is the whole difference. Everything else — the charts, the exports, the mobile app — is packaging.

Practically, this means three capabilities have to sit together in one place:

• Measurement at a resolution that matches the decision. Interval data at the circuit or asset level, not a single main meter read once a month.
• Continuous analysis that runs whether or not anyone is looking. Baselines, thresholds, and comparisons that update on their own.
• Delivery to a person who can act, with enough context that they know what to do without opening four other systems.

Break any one of those and you are back to reporting.

Three tests for real intelligence

If your current system fails any of these, you are reporting on history rather than operating your facility.

• Can you see the cost of a load in the last hour, not the last quarter? Cost, not just kilowatt-hours. A number in dollars is the only version that travels to finance without translation.
• Does the system tell you when something changed, or do you have to go looking? Anything that depends on someone remembering to check a dashboard will eventually be missed, usually during the week it mattered.
• Can you trace any number on a dashboard back to the raw interval that produced it? If a figure cannot be defended down to the underlying measurement, it will not survive its first challenge from a tenant, an auditor or a plant manager.

A fourth test is worth adding once the first three pass: can a new maintenance tech understand the screen without training? Intelligence that requires an expert interpreter is a consulting engagement, not a system.

What it is not

It is not a BAS. A building automation system controls equipment; it is very good at that and generally poor at measuring cost, trending power quality, or explaining why a bill moved. The two are complements. A BAS knows a valve position. It does not know that valve cost you $4,200 last month.

It is not a utility bill portal. Bill analytics can tell you the rate structure hurt you. They cannot tell you which of your 40 loads caused the 15-minute interval that set the peak.

It is not a data lake. Storing every point at high resolution is necessary but not sufficient. Plenty of facilities are paying to warehouse data nobody has ever queried. Storage without analysis is a cost centre with a dashboard bolted on.

It is not a one-off audit. An audit is a snapshot by a competent stranger. It ages immediately. Facility intelligence is the audit running continuously, which is why it catches drift — the slow return of the waste the audit eliminated.

Where equipment health fits

Energy and reliability are usually treated as separate programs with separate budgets, which is odd, because they read the same signal. A motor drawing more current than its baseline for the same output is both an energy problem and a mechanical one. Rising harmonic distortion is a compliance question and a heat-damage question. A compressor cycling more often is wasting power and wearing itself out.

Real facility intelligence treats electrical data as condition monitoring. The same interval stream that produces the cost report also produces the early warning, which is what makes the business case work: you are not funding two systems, you are funding one measurement layer that pays back twice.

Start where the money is

The fastest path to value is rarely a full retrofit. Meter the largest and most variable loads first — compressed air, refrigeration, chillers, major motor groups — then expand once the savings are funding the next phase.

A workable sequence:

1. Main service and the top five loads. This alone usually explains 70–80% of consumption and every demand peak.
2. Set baselines before you change anything. Two to four weeks of normal operation gives you the comparison you will need to prove any later claim.
3. Turn on a small number of alerts. Off-hours runtime, demand approaching the peak, and any load more than a set percentage above baseline. Three good rules beat thirty noisy ones.
4. Give the numbers an owner. An alert with no name attached is a notification, not a process.
5. Expand into the long tail once the first phase is demonstrably paying for itself.

The honest summary

Facility intelligence is not a product category so much as a standard: measurement fine enough to act on, analysis that never stops, and a path from a number to a person who owns it. If a system meets that standard, the label does not matter. If it does not, no amount of charting will make it an operating tool.