AIR (Asset Intelligence Rating) is meant to be interrogated. This page explains how the score is built — the scale, the four dimensions, how findings roll up, and how we keep buildings comparable. The full technical methodology is available to evaluators on request.
AIR rates how well a building is operating against what's achievable for a building of its kind — not how it was designed, and not its raw energy use. Higher is better. The bands below are guidance, not hard cutoffs; what matters is the findings behind the number and the direction it's moving.
0–59
Significant opportunity
60–74
Room to improve
75–89
Strong
90–100
Excellent
Each dimension is computed from your trend data and is itself fully traceable to the findings beneath it.
AIR dimensions · inputs
| Dimension | What it measures | Primary inputs |
|---|---|---|
| Energy efficiency | How much energy spend does useful work vs. leaks through faults | Simultaneous heating/cooling, economizer behavior, runtime, reset effectiveness |
| Occupant comfort | How consistently zones hold their comfort bands | Zone temp vs. setpoint, recovery times, deviation frequency and duration |
| Maintenance risk | Early behavioral signatures of failing equipment | Short-cycling, hunting, drift, abnormal runtime and start patterns |
| Controls health | Whether sequences still perform as designed | Reset engagement, loop stability, schedule adherence, setpoint integrity |
The exact weighting of dimensions into the composite is calibrated by building type and shared, in full, with technical evaluators.
Within each dimension, findings are weighted by their severity — verified dollar impact for energy, deviation magnitude and duration for comfort, and failure-risk for maintenance. A handful of expensive faults move the score more than a long tail of trivial ones, because that's where your attention should go.
A 200,000 sq ft school and a 600,000 sq ft research lab have very different baselines. AIR normalizes for building size and type so the score reflects operational performance relative to what's achievable for that class of building — which is what makes ranking a whole portfolio meaningful.
AIR is built on the same Prescriptiv fault-detection engine that grew out of Carnegie Mellon research, and the findings that drive the score are reviewed by LeanFM engineers who do commissioning before they reach you. The methodology evolves as deployments surface new failure modes — and every change is reflected the next time your building is analyzed.
For technical evaluators: we share the full methodology — dimension weightings, normalization approach, and detection logic — under NDA. Request it here.
Request a sample analysis and we'll compute your building's first AIR score — with every rating opened to its evidence.