IMAS Analytics.AI
HomeSolutionsPeople counting & visitor analyticsXPERIO dashboardVisitor forecastingAI building energy managementIndustriesShopping mallsRetail chainsAirports & transport hubsOffice & commercial buildingsTry itProjectsInsightsAbout usPartnersContact us

DATA INTELLIGENCE

What Is a Smart Building? How Smart Building Systems Work, and Why Occupancy Data Is the Piece Most Often Missing

A smart building is a building that connects its systems through data, so they run according to how the building is actually used, not according to a schedule set once and left alone. It is not measured by how many devices are installed. It is measured by whether data from one system actually helps other systems, and helps people make decisions. Most buildings know fairly well what the temperature is and whether their equipment is running. What they usually do not know is where people are right now, and how many. This article explains the layers of a smart building system, what each layer does, and how occupancy data changes the way air conditioning, lighting, cleaning, space planning and security work.

High-rise building in a business district at dusk with some floors lit and others dark

What Kind of Building Counts as a Smart Building?

A practical definition for building managers is a building that can do these 4 things continuously:

  1. Measure collect data from equipment, the indoor environment and how people use the building
  2. Connect let data from one system reach another, instead of staying locked inside one system
  3. Decide use that data to adjust how systems run, with people setting the rules and limits
  4. Check look back at whether the change worked, then adjust again

A building with a modern Building Management System that still switches everything on and off on the same fixed schedule every day is not yet a smart building in this sense, because its data is not used to make decisions. An older building with no central control system, whose engineering team uses meter data and occupancy data to plan the week, has already started down this path.

How Many Layers Does a Smart Building System Have?

A smart building system can be split into 5 layers. Each layer depends on the one below it.

Layer What it includes What it does
1. Field devices Temperature, humidity and carbon dioxide sensors, electricity meters, people-counting sensors, valves, dampers, switches Measure values and receive commands to open, close or adjust
2. Control systems Building management system, lighting control, access control, lift control, fire alarm system Operate equipment according to set values
3. Data connectivity Building network, gateways, standard protocols such as BACnet, Modbus and MQTT, and API connections Move data between systems and out to platforms
4. Analytics platforms Energy management system, occupancy analytics platform, dashboards, forecasting Combine data from several sources, find anomalies, point to what should change
5. People and processes Building engineers, cleaning team, security team, building management, work-order system Decide, act and follow up

Layer 5 is often forgotten when people talk about smart buildings, but it is the layer that turns data into results. If a dashboard shows a floor running full air conditioning on a day with no one there, and no one picks up the issue, that data is worth nothing.

The difference between an energy management system (EMS), a building energy management system (BEMS) and a building management system (BMS) is explained in the article What Is an EMS in a Building, and How Does It Differ from BEMS and BMS?

A Building Needs to "Know" 3 Things, but Most Know Only 2

Almost every system in a building exists to look after the people inside. A building that wants to adjust well needs to know 3 things.

What it needs to know Example data Do most buildings have it?
Environment Temperature, humidity, carbon dioxide, daylight Usually, from the building management system
Equipment Equipment status, electricity use, running hours, alarms Usually, from the building management system and meters
People How many people, which zone, when they arrive, when they leave Usually not, or only as an estimate

Without people data, systems run on assumptions instead, such as the design occupancy, opening hours in the lease, or the team's experience. These assumptions may have been right on the day the building was designed, but building use keeps changing: hybrid work, changing tenants, and shopping-mall visitor patterns that differ sharply between weekdays and weekends.

Why the Proxy Data You Already Have Is Not Enough

Many buildings have some data that looks like it could stand in for occupancy, but each type has limits.

Proxy data What it can tell you Limits
Access cards How many times cards were tapped to enter Does not show when people leave, misses people who follow others through, visitors and shoppers have no card
Motion sensors Whether someone is moving in a room Shows only presence or absence, not how many, may miss people sitting still
Carbon dioxide levels Whether a room is filling up or emptying Responds slowly, depends on ventilation and room size
Wi-Fi signals from phones Number of connected devices across a wide area Counts only people with devices switched on, some people carry several, and needs care with personal data

This data is still useful. But if you need a reliable answer to "how many people are in this zone right now", you need sensors built for counting people. How "is anyone there" sensors differ from "how many" sensors is explained in the article What Is an Occupancy Sensor?

Which Building Systems Change When You Know Occupancy?

This table summarises what each system in a smart building runs on today, and what occupancy data changes.

System Runs on today With occupancy data Who decides
Air conditioning Fixed schedules and set temperatures Set start and stop times per zone to match real use, spot zones running at full load with no one there Building engineers
Ventilation Outdoor air volume based on design occupancy Adjust outdoor air to actual occupancy in spaces where headcount varies widely Building engineers, within indoor air quality requirements
Lighting Schedules, or motion sensors room by room Schedule common-area lighting to occupied hours, reduce lighting in closed zones Building engineers, building management
Lifts and escalators A fixed number in service by time of day Adjust how many are in service to the flow of people in each period Building management
Cleaning Fixed rounds for every area More rounds for restrooms and busy areas, fewer for areas that are barely used Cleaning supervisor
Space planning Headcount on the payroll, or how users feel See actual use of floors, zones and meeting rooms to support space planning Space management, executives
Safety and emergencies Patrols and CCTV Know how many people are in the building and each zone right now, spot unusual crowding Security team, following building procedures
Service and queues Supervisors' experience See queues forming and staff the busiest periods Service supervisors
Overall energy management Energy per floor area Track energy per building user alongside it, compare busy days with quiet days Energy team, executives

Two points stand out from this table:

  • Occupancy data is shared data. Collected once, it serves several teams: engineering, cleaning, security and space management. Most sensors in a building serve only one system.
  • Every row still has a person deciding. Data gives a clearer picture, but changing building settings must take into account occupant comfort, air quality, equipment limits and tenant agreements.

Air Conditioning and Ventilation

Air conditioning is a major energy user in most commercial buildings in Thailand, and it is the system that benefits most clearly from occupancy data, for 2 reasons. First, people are a source of heat inside the space. Second, the amount of outdoor air that must be brought in is calculated from both the number of people and the floor area, following ventilation standards such as ASHRAE Standard 62.1. Outdoor air in Thailand is hot and humid, so bringing in more air than the actual occupancy needs adds to the cooling load.

In shopping malls this applies to the hours before opening and after closing, zones that are quiet at certain times of day, and running chillers to match the forecast load. These are covered in detail in the article Mall Electricity Costs: Where Can You Cut When You Know Real Visitor Numbers? For office buildings, see how to set schedules floor by floor in the article What Is an Occupancy Sensor?

Building Services and Space Planning

Cleaning, reception and security are major staffing costs for a building, and are usually organised around fixed rounds. Occupancy data helps supervisors send people where the space is actually being used, such as restrooms on a floor with a large meeting room on an event day, or seating areas in a mall on a weekend afternoon. With several months of history, space managers can also see which floors or rooms are underused for their size, and use that to support space planning.

Safety and Crowd Management

Knowing how many people are in each zone right now supports emergency management and evacuation drills. It is used alongside the building's own headcount procedures, not instead of them. In areas with large crowds, such as a mall event space or a screening point in a passenger terminal, seeing density and queues as they form lets staff open more service points or manage walkways before overcrowding sets in. Read more in the article Managing Queues and Passenger Flow at Airports and Transport Stations

Levels of "Smart": You Do Not Need to Start with Automatic Control

Many projects stall because they aim for systems to control each other automatically from day one, before the team has data it trusts. Thinking in levels makes it possible to actually start.

Level What the system does What people do
1. See Show occupancy and energy use per zone on the same dashboard Read the data, look for mismatches
2. Alert Send alerts when something is abnormal, such as a zone with no one in it but systems at full load Take the issue, check on site, fix it
3. Recommend Suggest changes, such as a new schedule based on history and forecasts Review, approve or reject
4. Control within rules Adjust system settings automatically within limits people have set Set the rules and limits, review results regularly

Most buildings gain a lot from levels 1 and 2 first, because the team starts to trust the data and learns which rules suit its own building before moving to automatic control. At every level, artificial intelligence (AI) is a tool that helps read data and suggest options. People on the building team still make the decisions.

How to Start Making an Existing Building Smarter

  1. Survey what you already have. Building management system, main meters and sub-meters, lighting control, access control. Which systems can export data, and through which channels?
  2. Pick 1–2 problems to solve. For example, setting air conditioning schedules floor by floor to match real use, or scheduling restroom cleaning by actual use.
  3. Define zones that match control zones. This matters and is often missed. If occupancy will be used to adjust air conditioning, the counting zones should match the zones the air conditioning can control separately, not the plan the leasing team uses.
  4. Place sensors to cover every entrance and exit of each zone. If one entrance is not counted, the occupancy figure for the zone drifts over time.
  5. Connect the data. Send occupancy data into the same platform as energy data, or to other systems through an API, as the building needs.
  6. Measure before and after. Record data before changing a schedule or rule, then compare with the period after, taking differences in weather and occupancy into account.
  7. Assign owners. Who watches the dashboard, who receives alerts, and who approves changes to system settings.

Occupancy Data and Privacy

A smart building does not need to know who the people inside are. What building systems need is overall numbers and locations. The Xovis 3D sensors that IMAS uses process data inside the sensor and export only numerical data, such as counts and movement coordinates, with no biometric or personal data. This design reduces the data-protection burden, but the building should still tell occupants why people are being counted. Read more in the article People Counting and the PDPA, which covers Thailand's Personal Data Protection Act B.E. 2562 (2019).

How IMAS Adds Occupancy Data to a Smart Building

Cleaning cart being pushed along a corridor outside office restrooms

IMAS offers both people counting and building energy management, so it can supply the missing "people" piece for buildings that already have control systems.

  • 3D people counting counts entries and exits by direction in real time, shows how many people are in a space and where crowding occurs, and is designed to count without identifying individuals. The sensors can export data over several channels, such as HTTP(S) and MQTT(S), and provide a documented API.
  • XPERIO dashboard shows occupancy and dwell time, compares daily to yearly, and sends data to other systems through a REST API and webhooks. Integration with each building's systems uses existing connectors where available, or is delivered as a project for that building.
  • Visitor forecasting predicts visitor numbers ahead of time, to plan system schedules and staffing.
  • AI building energy management with Saturn Pyro on the IoTWatt 4.0 platform connects the meters, building management system and IoT devices a building already has, alerts on abnormal energy use, and turns findings into corrective tasks with named owners.

The systems show how the building is really used and where to look first. Decisions on adjusting systems stay with the building engineers and managers who know the building's limits best. See the overview for these buildings on the office and commercial building solutions page and the shopping mall solutions page.

Frequently asked questions

−What is a smart building?

A building that connects systems such as air conditioning, lighting, lifts, security and people counting through data, so they run according to how the building is actually used, and help the building team make better decisions.

+How is a smart building different from a building with a building management system (BMS)?

The building management system is one layer of a smart building. It operates equipment according to set values. A building becomes smart when data from several systems is combined, analysed and used to keep adjusting how it runs.

+Do we need to replace all our existing systems?

No. Most buildings start from the systems they have, then add the data they are missing, such as sub-meters or occupancy data, and bring the data together on one platform.

+What is occupancy data?

Data on how many people are in a given space at a given time, for example the number of people on each floor every 15 minutes. It is used to set system schedules, organise building services and plan space.

+Can we use access-card data instead of counting people?

It can help, but access-card data shows how many times people tapped in. It does not show when they leave, misses people who follow others through, and does not count visitors or shoppers without cards, so it is not suitable when you need the number of people in a zone right now.

+Does occupancy data in a smart building require cameras that see faces?

No. 3D sensors process data inside the device and export anonymous numbers. Building systems need overall counts and locations, not who is who.

Want to Know Where Your Building Is Missing People Data?

The IMAS team helps survey your existing systems, define zones that match control zones, and plan how to connect occupancy data with energy data. Details and costs are provided in the quotation.

Book a system demo · See AI building energy management

References

  • Xovis AG. Building management solutions. Website xovis.com
  • Xovis AG. Technical Datasheet PF-L and PC2SE (data output channels and API). Website xovis.com
  • CountMatters AS. XPERIO analytics platform and People Counting (REST API and webhooks). Website countmatters.com
  • Saturn Pyro Sdn Bhd. IoTWatt 4.0 Digital Energy Audit as a Service. Website saturnpyro.com
  • ASHRAE. Standard 62.1 Ventilation and Acceptable Indoor Air Quality (outdoor air calculated from occupancy and floor area)