Occupancy sensors detect whether people are present in a space. The building then uses that signal to control lighting, heating, cooling and ventilation. In a conventional setup, each sensor triggers one local action, such as switching a light off when a room empties. In agentic building operations, occupancy data becomes a live input for AI agents. The agents adjust HVAC, release unused rooms, schedule cleaning and flag floors worth shutting down, continuously and across the whole portfolio.
This guide covers how occupancy sensors work, which type to use where, how to place them, and how to get more from the data than a light switch.
How do occupancy sensors work?
An occupancy sensor watches a space using infrared, sound waves, microwaves or other signals. When it detects that the space has changed from occupied to vacant, or the reverse, it sends a signal to a controller. That controller might switch the lighting or change the HVAC mode for the zone. Most sensors also apply a time delay, keeping the space “occupied” for a set period after the last detection, so lights don’t switch off on someone sitting still.
Types of occupancy sensors
Passive infrared (PIR). PIR sensors detect the heat signature of people moving across their field of view. They are inexpensive and reliable in enclosed spaces with clear sightlines, such as private offices and corridors. They need line of sight, though, and can miss small movements like typing.
Ultrasonic. Ultrasonic sensors emit high-frequency sound and measure changes in its reflection. They are more sensitive to small movements and work around partitions, which suits restrooms, open offices and spaces with obstructions. They are also more prone to false triggers from air movement.
Microwave. Microwave sensors emit microwave signals and measure reflections. They cover large areas, but because the signal can pass through thin walls and glass, they may detect movement outside the intended zone. They need careful sensitivity tuning.
Dual-technology. Dual-technology sensors combine two methods, typically PIR with ultrasonic or microwave. They report occupancy only when both agree, or keep it only while either one still detects presence. This reduces false on and false off events, which makes them a common choice for open-plan offices and conference rooms.
People counters and presence sensors. Newer devices such as thermopile arrays, time-of-flight sensors, radar and anonymized optical counters report how many people are present, not just whether anyone is. That count is what space utilization and demand-controlled ventilation need. Because these devices can raise privacy questions, see our guide to GDPR-compliant presence sensors.
Where should occupancy sensors be placed?
Placement decides whether a sensor saves energy or annoys occupants. Six factors matter most.
Room size and shape. Large or L-shaped rooms need several sensors, or one technology that can see around corners.
Mounting height. Height changes coverage. Follow the manufacturer’s coverage pattern for the actual ceiling height.
Line of sight. PIR sensors need an unobstructed view of where people actually sit, not just the doorway.
Interference. Keep sensors away from HVAC diffusers, heat sources and reflective surfaces, which cause false readings.
Room function. Match the sensor type to the use. Examples are ultrasonic or dual-technology in restrooms, and dual-technology where people sit still for long periods.
Data needs. If the data will feed space analytics or ventilation control, not just lighting, choose sensors that report counts and can be integrated through an API.
How far do occupancy sensors reach?
Range depends on technology, mounting height and model. Typical wall- and ceiling-mounted sensors cover roughly 3–15 metres (10–50 feet). Microwave sensors can reach further, which is useful in large open spaces but makes them harder to confine to one zone. Always design to the manufacturer’s coverage pattern rather than a headline range.
Occupancy sensor vs vacancy sensor vs motion sensor
Occupancy sensors turn systems on automatically when someone enters and off automatically after the space empties.
Vacancy sensors require the occupant to switch lights on manually, then turn them off automatically once the space is empty. Because nothing switches on by accident, they often save more energy, and some energy codes require them in certain spaces. The Lighting Controls Association explains the difference in detail.
Motion sensors is the broad term for any device that detects movement and triggers an event, whether that event is a light, an alarm or a camera. Every occupancy sensor uses some form of motion or presence detection, but not every motion sensor is used to manage occupancy.
From local automation to agents that act on occupancy data
A sensor wired to one light fixture is automation. It follows a single fixed rule in a single room. Most of the value in occupancy data comes when it is connected across systems and used to make decisions. In agentic building operations, occupancy works as a signal that AI agents act on:
HVAC and ventilation. Agents condition and ventilate spaces according to how many people are actually in them. They reduce output in empty zones and ramp up ahead of expected arrivals. This links directly to the indoor air quality work of an environmental control system and to the energy savings of an energy management control system.
Room release. Agents free up meeting rooms that were booked but never used.
Floor shutdown. Agents put an entire floor into setback when it is empty, instead of conditioning a half-empty building.
Needs-based cleaning. Agents trigger cleaning only where spaces were used, rather than following a fixed schedule.
Sensor health. Data quality agents spot sensors that have gone silent, drifted or report impossible values, and raise a work order before bad data drives bad decisions.
Agents act within guardrails. For example, a missing occupancy reading should never cut ventilation below minimum outdoor air requirements. The hard limits are enforced by the platform and the BMS. For how these limits are set, see permission policies and guardrails for AI agents in real estate.
For the full path from raw presence data to agents that act on it, including why occupancy is not the same as utilization, read From sensors to agents: what occupancy data can do.
Getting occupancy data ready for agents
Buildings often have more occupancy signal than they realize. CO₂ levels, badge events, Wi-Fi counts and booking systems all correlate with presence. Three things turn that signal into something agents can use.
Connectivity. Sensors and building systems need to be reachable through APIs, via connectors and partner integrations.
A shared data model. RealEstateCore gives each sensor context: which room it sits in, how big that room is, and which air handling unit serves it. That context is what turns a presence reading into a utilization figure. It also means you can replace a sensor vendor without losing your history.
Permissions. You decide what agents may adjust based on occupancy, and what needs a human to approve.
AI agent-ready buildings: onboarding explained walks through the process.
Why occupancy sensors matter for commercial real estate
Lower energy use. Systems that only run when spaces are used waste less energy. When agents act on occupancy continuously across HVAC and lighting, the savings compound. ProptechOS deployments average around 30% energy savings, and Vasakronan cut heating and cooling energy use by 36%.
Better tenant experience. Rooms are comfortable when people arrive, meeting rooms are actually available, and air quality keeps pace with how busy a space is. See tenant experience and EU air quality compliance.
Smarter space decisions. Utilization data shows which floors, rooms and desk types are underused. That evidence can inform lease, consolidation and renovation decisions. See combining building and workplace data to unlock office insights.
Evidence for ESG. Occupancy-adjusted energy data gives a more honest picture of building performance for sustainability and compliance reporting.
Occupancy-driven operations with ProptechOS
ProptechOS connects occupancy sensors, booking systems and your existing BMS in one open data model built on RealEstateCore. It then runs AI agents that act on the data, adjusting HVAC, releasing rooms, flagging empty floors and monitoring sensor health. Every agent works within scoped permissions and full audit logs.
See how ProptechOS works, read what agentic means in building operations, or book a demo. You can also start a free trial.
FAQ
How do occupancy sensors work?
They detect people using infrared, ultrasonic, microwave or other signals and send an occupied or vacant status to a controller. The controller then adjusts lighting, heating, cooling or ventilation.
Which type of occupancy sensor is best for an office?
PIR suits enclosed offices with clear sightlines. Ultrasonic or dual-technology sensors suit open-plan areas, restrooms and rooms where people sit still. People counters are the best choice when the data will drive ventilation or space utilization decisions.
What is the difference between an occupancy sensor and a vacancy sensor?
An occupancy sensor switches systems on and off automatically. A vacancy sensor requires a manual switch-on and only switches off automatically, which avoids unnecessary activations.
How do AI agents use occupancy data?
Agents use live occupancy to adjust HVAC and ventilation, release unused meeting rooms, put empty floors into setback, trigger cleaning where spaces were used, and flag faulty sensors. They do this continuously and within limits set by the operations team.
Do I need new sensors to start?
Often not. CO₂ sensors, badge systems, Wi-Fi counts and booking data already correlate with occupancy. Connected into one data model, they can estimate occupancy where no dedicated presence sensor exists.
