Workplace Comfort Optimization with IoT Sensors
February 25, 2026 · Sarah Okafor
An office is a machine for converting salaries into output, and its performance depends on an environment people stop noticing the moment it is right — and notice the moment it is wrong. Thermal discomfort, stale air, and glare are not minor irritations: the field evidence ties them to measurable drops in task performance, and because labor cost dominates an office’s P&L by orders of magnitude, a single-digit productivity swing is worth more than the entire energy bill. That asymmetry is the business case — the sensor layer that feeds the HVAC system also feeds the human dimension.
The Comfort Equation
Human thermal comfort is not simply air temperature. It balances six variables — air temperature, radiant temperature, humidity, air speed, clothing insulation, and metabolic rate — evaluated with the predicted mean vote (PMV) and predicted percentage of dissatisfied (PPD) indices of ASHRAE Standard 55. The uncomfortable truth: no single temperature satisfies everyone, and up to 20 percent of occupants may be dissatisfied by design.
The practical consequence: comfort optimization is not “set the thermostat to 22 °C.” It is controlling operative temperature — the combined effect of air and radiant temperature, the variable a wall thermostat cannot see. A room with a cold glass facade can read 22 °C while occupants near the window feel cool, because the radiant temperature at their position is lower. Room-level sensors measuring radiant and operative temperature reveal the gradients fixed thermostats miss.
Why the Same Setpoint Feels Different
Two adjacent floors at identical setpoints can feel entirely different — the cause is almost never the thermostat. Cold windows or hot ceilings drive discomfort even at a “correct” air temperature — why radiant systems win comfort complaints: they fix the radiant component. A VAV box dumping air at 0.25 m/s across a desk reads as “too cold” even at a warm setpoint. And thermal mass plus solar gain make operative temperature lag the thermostat by hours — a floor cold at 9 a.m. and warm at 3 p.m. is a scheduling problem, not a setpoint problem.
Carbon Dioxide: The Ventilation Signal
Carbon dioxide is not a health hazard at office concentrations, but it is a cheap, excellent proxy for whether ventilation is flushing the space. Outdoor air runs around 400 ppm; a well-ventilated office holds 500 to 700 ppm; a crowded one drifts toward 1,000 ppm. The replicated studies consistently link elevated CO₂ with reduced cognitive performance — marked drops in decision-making at 1,000 ppm versus 600 ppm.
Ventilation optimization has a dual payoff. Demand-controlled ventilation — modulating outdoor-air dampers with CO₂ instead of running at a fixed design minimum — improves air quality when occupied and stops over-ventilating when empty. A conference room empty 80 percent of the day does not need its design rate at 3 a.m.
Humidity and the Comfort Band
Relative humidity operates within a narrow comfort and health band. Below about 30 percent RH, occupants get dry eyes, dry skin, and static discharge; above about 60 percent, the space feels stuffy and microbial growth risk rises — which is why many organizations target 40 to 60 percent RH through winter.
Humidity is also an energy signal: a humid climate’s cooling load is dominated by latent dehumidification, while a dry climate’s heating season drives humidification energy. Per-zone RH monitoring exposes the lobby over-dehumidifying on a faulty valve or the corner office running dry beyond the diffuser’s reach.
Lighting and the Circadian Dimension
Comfort extends to light. A space held at a uniform 300 lux and constant color temperature feels flatter than one that tracks the day. Circadian lighting research — higher melanopic light in the morning, warmer tones toward evening — shows modest but consistent gains in alertness and sleep, and glare and flicker remain the fastest routes to the complaint desk. Sensor-verified illuminance and CCT tuning fold light into the comfort program without a separate retrofit.
The Measurement Stack
A practical comfort layer is modest in hardware and large in signal:
| Parameter | Sensor | What it drives |
|---|---|---|
| Operative / radiant temperature | Globe thermometer, radiant sensor | Zonal setpoint reset, radiant complaint diagnosis |
| Air temperature | Room sensor | HVAC staging |
| Relative humidity | Room sensor | Humidification/dehumidification control |
| CO₂ | NDIR sensor | Demand-controlled ventilation |
| Illuminance | Photosensor | Lighting levels, glare detection |
| Occupancy | People counting, access control | Setpoint scheduling, ventilation staging |
The engineering decisions that matter are density and placement: one sensor per zone is the floor; workstation-level sensors reflect what people experience; and calibration keeps a drifted CO₂ sensor from corrupting the ventilation logic.
From Data to Action
The sensor layer pays only when it changes the control strategy. The highest-value actions:
- Zone-level setpoint reset driven by operative temperature and occupancy, so empty or comfortable zones do not keep conditioning.
- Demand-controlled ventilation so air quality is guaranteed when occupied and energy is not wasted when empty.
- Draft and asymmetry remediation — the sensor identifies the cold corner or dumping VAV box, and the fix (diffuser, rebalance, radiant panel) is targeted rather than guessed.
- A comfort-and-energy review loop: weekly, compare comfort metrics against energy intensity per floor and flag the floor that is both uncomfortable and wasteful — usually one root cause.
A Worked Example
A 200-person floor runs ventilation at a fixed 800 cfm design rate from 6 a.m. to 8 p.m., every zone held to a single 22 °C setpoint. CO₂-driven demand control cuts average ventilation flow 35 percent while keeping peak CO₂ under 800 ppm. The same data shows the south wing drifting to 24 °C every afternoon under solar gain; a zone-level reset cuts that wing’s cooling demand and its complaints together.
The counterfactual makes it worth it: a one-degree deviation from the comfort optimum is associated with roughly a 2 percent performance drop — about $280,000 a year on a 200-person floor at $70,000 loaded cost per employee, two orders of magnitude above the energy bill. Comfort optimization is not an energy expense; it is the most profitable control loop in the building.
Workplace comfort optimization is the rare project where the technical and human cases are the same: the sensor layer that protects the employee also exposes the building’s waste. Integrar IoT places the comfort sensor streams on the same BACnet and MQTT backbone as energy, access control, and lighting — so the office is managed as one comfort-and-cost system, not a set of separately argued thermostats.