Your Wi-Fi Could Soon Know When Someone Is in the Room

Imagine a small office that wants to know whether a meeting room is genuinely being used. Today, it might use a booking calendar, a door sensor, an infrared sensor or a camera. Another possibility is to use the Wi-Fi already moving through the room and look for changes in the radio environment when people enter, leave or move around.

That idea sounds futuristic because Wi-Fi is normally described as a way to connect laptops and phones. But radio waves do not simply travel from an access point to a device and disappear. They reflect, scatter, weaken and take different paths through a physical space. People, furniture, doors and other objects affect those paths.

The important question for a business is not “Can Wi-Fi see through walls?” It is “What limited, tested inference can this particular system make in this particular space, and what will we do with that information?”

What Wi-Fi sensing actually measures

A wireless link has a transmitter and a receiver. The receiver can observe properties of the signal it receives, such as strength, timing, phase or how different parts of the signal behave across the channel. Different Wi-Fi generations and chipsets expose different measurements, and a product may process them internally rather than making them available to an ordinary network administrator.

When somebody moves between, near or around the radio paths, the measured signal can change. A body may absorb some energy, reflect some energy and alter the route taken by other parts of the signal. A chair being moved, a door opening or a group of people entering can also change the pattern.

The software is not looking at a photograph hidden inside the Wi-Fi. It is comparing measurements with a model, baseline or threshold. That is why room layout, access-point position, furniture, radio interference and the number of people all matter.

How the inference is made

A simplified sensing pipeline looks like this:

RADIO EXCHANGE → MEASUREMENTS → CLEAN-UP → MODEL OR RULE → EVENT

1. 1Radio exchange: one or more Wi-Fi devices transmit and receive signals in the area.
2. 2Measurements: the system records permitted characteristics of what the receiver observed.
3. 3Clean-up: software filters noise, deals with missing samples and establishes a baseline for the room.
4. 4Model or rule: an algorithm compares new measurements with the baseline or a trained pattern.
5. 5Event: the system produces a limited result such as movement, occupancy, zone change or an alert.

A simple system may only report “motion” when a signal changes beyond a threshold. A more advanced system may combine many radio paths and a machine-learning model to estimate occupancy or an activity. The second approach can be more capable, but it can also be harder to explain, tune and validate.

Why IEEE 802.11bf matters

IEEE 802.11bf is the wireless LAN sensing amendment. The IEEE Standards Association lists IEEE 802.11bf-2025 as an active standard, published on 26 September 2025. It defines changes to Wi-Fi MAC and PHY operation to enhance wireless LAN sensing across supported licence-exempt frequency ranges.

That matters because a common standard can make it easier for equipment and software from different parts of the ecosystem to support sensing in a more predictable way. It does not mean that every access point has been upgraded, every client exposes the required measurements or every vendor feature will interoperate automatically.

The standard is also not a promise of a particular application. It does not guarantee that an office product will count people accurately, that a home router will detect a fall or that a system can identify an individual. Those outcomes depend on the implementation, radio environment, data processing and testing behind the product.

IEEE Standards Association: IEEE 802.11bf-2025 →

IEEE 802.11 Working Group: WLAN Sensing task group →

This is more than a thought experiment

Wi-Fi sensing has been the subject of research for years, including work on presence, occupancy, gesture, activity and environmental monitoring. Technical demonstrations and prototypes have used measurements from commercial wireless hardware, while network and building-technology suppliers have described sensing or occupancy products that combine Wi-Fi telemetry with other inputs.

That evidence shows that the underlying idea is real. It does not show that a particular business can enable reliable sensing by changing one setting on an existing router. A research result may use carefully placed radios, a known room, controlled training data and a narrow task. A commercial deployment has to cope with changing furniture, different bodies, pets, interference, device updates and people using the space in unexpected ways.

IEEE technical overview of the 802.11bf sensing amendment →

Nokia Bell Labs: Network as a Sensor →

Cisco: Emerging trends in Wi-Fi sensing →

What could a business use it for?

The most credible early uses are usually about spaces and events rather than identifying people. That makes Wi-Fi sensing potentially interesting for organisations that want a lower-friction way to understand how rooms or buildings are used.

1. Occupancy and room use

A system might estimate whether a meeting room, reception area, classroom, shop floor or shared workspace is occupied. That could help a business compare bookings with actual use, reduce wasted space and understand which areas need more capacity.

The result should be treated as an estimate with known accuracy, not an unquestionable headcount. A room with two people may produce a different signal pattern from the same room with six people, but the relationship is not automatically universal.

2. Unexpected movement

Outside normal hours, a sensing system could raise an alert when movement occurs in a restricted office, stockroom or plant area. This might complement an alarm, access-control log or physical check. It should not be described as a complete security system: radio sensing can produce false positives and may not explain the cause of an event.

3. Lighting, heating and ventilation

Presence information could help control lights, heating, cooling or ventilation so that a space is conditioned when it is actually in use. In principle, this could reduce energy waste while making a building more comfortable.

Automation needs a useful delay and an override. A light that turns off whenever somebody sits still, or heating that reacts to a single noisy event, will frustrate people. Measure comfort and energy outcomes rather than assuming a sensor automatically creates savings.

4. Building utilisation

A multi-room business could use aggregated occupancy patterns to understand when floors, desks or facilities are genuinely used. This might inform an office move, cleaning schedules, maintenance planning, space booking or a decision to consolidate a site.

5. Assisted living and welfare support

A camera-free signal may be appropriate in a bedroom, bathroom or supported-living space where a camera would be intrusive. Possible uses include detecting an unusual lack of movement, a fall-like pattern or a change in normal routine that prompts a human welfare check.

This is a sensitive scenario. The system may miss an event, infer the wrong event or fail when the network or power is unavailable. It should support a human-led care plan, not silently replace one. People need to understand what is being monitored, who receives an alert, how long records exist and what happens when the technology is uncertain.

What Wi-Fi sensing cannot promise

The phrase “Wi-Fi can see through walls” makes a striking headline, but it hides the practical limits. A sensing system is making an inference from a radio environment, and radio environments are messy.

Other moving things matter too. Fans, doors, lifts, pets, stock, partitions, water, machinery and neighbouring networks can alter radio measurements. A system trained in an empty demonstration room may behave differently when a real business is busy.

Camera-free does not mean data-free

It is tempting to describe Wi-Fi sensing as private because it does not produce video. That is too simple. The system may collect raw or processed signal measurements, create occupancy events, infer activity, send alerts and retain a history of when an area appeared occupied.

The privacy impact depends on the whole design. A local device that discards raw measurements and reports only a coarse room-level state is different from a cloud service that keeps detailed time-stamped events, combines them with named device identities and allows managers to review a person’s movements.

If sensing is used to monitor staff, attendance, productivity or behaviour, the employment and data-protection implications become more serious. “We are not filming anyone” is not a sufficient explanation for workplace monitoring.

A practical UK privacy starting point

A business should define the purpose before choosing the technology. “We want to use AI and Wi-Fi to understand the building” is not a useful purpose. “We want to reduce heating in unoccupied meeting rooms, using a room-level signal and deleting event data after a short period” is much clearer.

1. 1Describe the outcome you need and test whether a less intrusive sensor or process could achieve it.
2. 2Map the data: radio measurements, derived features, alerts, occupancy history, device identifiers and any links to people.
3. 3Decide who is controller and who is supplier or processor for each part of the service.
4. 4Set a lawful basis, transparent notices, access restrictions and a retention period appropriate to the purpose.
5. 5Complete a data protection impact assessment where the processing is likely to create a high risk, and keep the reasoning on record.
6. 6Test accuracy across real layouts, times, people and edge cases before relying on alerts or decisions.
7. 7Keep a human review and a fallback process for safety, welfare and security scenarios.

The ICO describes DPIAs as an essential part of accountability and says they are required for processing likely to result in a high risk to people’s rights and freedoms. The right threshold depends on the proposed processing, not on whether a camera is present.

ICO: What is a data protection impact assessment? →

ICO: Data protection by design and by default →

Questions for the supplier

Before a purchase, ask for answers about the actual model and service you will receive, not just a brochure describing what Wi-Fi sensing might do in general.

• Which exact access points, radios, clients, firmware versions and regions are supported?
• Does the system use existing Wi-Fi traffic, dedicated sensing exchanges or additional hardware?
• Does it report movement, presence, a count, a zone, an activity or an identity?
• What accuracy and false-alarm rates were measured, in what type of building and with how many people?
• What changes when furniture, partitions, access-point positions or neighbouring networks change?
• Are raw measurements processed locally, sent to the supplier or stored in a cloud account?
• What event history, identifiers, logs and support data are retained, and for how long?
• Which supplier staff, administrators or subcontractors can access the data?
• Can the business export and delete its data, and can it disable sensing without losing ordinary Wi-Fi?
• What happens if the network, internet connection, power, firmware or cloud service fails?
• What security controls protect the sensing data and the management account?
• Will the supplier help with a DPIA, data map, privacy notice and data-processing terms?

A sensible pilot for a small business

Do not start by deploying sensing across the whole building. Choose one low-risk, well-understood space and define what success means before turning it on. Use the pilot checklist below to record the decision before you buy equipment or enable a feature.

Jump to the Wi-Fi sensing pilot checklist →

1. 1Choose a use case such as meeting-room utilisation, not employee scoring.
2. 2Write down the minimum result needed, for example occupied or not occupied in five-minute intervals.
3. 3Record what the system will collect, where it will process it and when it will delete it.
4. 4Test empty, occupied, still, crowded, rearranged and noisy conditions.
5. 5Compare the result with a temporary manual count or an appropriate non-identifying sensor.
6. 6Record missed events, false alarms, network failures and user feedback.
7. 7Review the privacy assessment and supplier answers before expanding the pilot.

Wi-Fi sensing pilot checklist

Work through these questions with the supplier and the people responsible for your network, facilities and privacy. A “no” or “not yet” is a reason to pause and clarify the pilot, not a problem to hide.

• Hardware and firmware: Confirm the exact access points, radios, client devices (if relevant), firmware, software plan and region are supported. Do not assume an ordinary router supports sensing.
• Measurement scope: Define one room or zone and the smallest useful result, such as occupied or not occupied. Exclude identity, productivity scoring and continuous person tracking unless there is a separately justified need.
• Data flow: Record which raw measurements, derived features, alerts, identifiers and locations are created, where they are processed, whether anything leaves the premises and which parts can be switched off.
• Accuracy test: Compare results with manual observations or an appropriate non-identifying reference across empty, occupied, still, crowded, rearranged and noisy conditions. Record false positives, missed events and the limits of the test.
• Fallback: Write down the human check or existing process used when the signal is wrong or the network, power, internet, firmware or cloud service fails. Do not make the pilot the only safety, welfare or security control.
• Retention: Set the shortest useful retention for raw measurements, derived data, event history and logs, then confirm that deletion can be carried out and checked.
• Access and security: Give access only to named roles, protect the management account and check supplier and subcontractor access. Do not give HR or managers access by default.
• Transparency and review: Tell affected people what is sensed, why, where, for how long and who can answer questions. Name the pilot owner, success and stop criteria, review date and triggers for pausing or reverting the change.

How this fits with your existing network

Wi-Fi sensing does not remove the need for a properly designed network. Coverage, access-point placement, channel planning, power, firmware, segmentation and monitoring still matter. A business that already has unreliable Wi-Fi should fix the network before adding a new inference layer to it.

Wi-Fi 7 for Business: Should You Upgrade Yet? →

Our Website Worked Everywhere Except the Office →

Business Broadband Contracts: What Should You Check? →

Treat the sensing service as another connected system. Give it only the network access it needs, protect its management accounts with strong authentication and include it in supplier, backup, incident and change records.

Zero Trust Security: What Does It Actually Mean? →

IT Club View

Wi-Fi sensing is interesting because it turns a familiar communications network into a possible source of environmental information. The strongest business cases are likely to begin with coarse, practical questions: is this room being used, is an area unexpectedly active, or can building services respond better to real occupancy?

The weakest business cases begin with the most ambitious promise: that an existing router can quietly understand people. A radio measurement is not a video feed, but it can still become a record about people. A machine-learning model can be impressive in a demonstration and unreliable in a changed room. A camera-free product can still be intrusive if nobody knows it is running or if managers use it to infer individual behaviour.

Our view is to start with the smallest useful inference, the shortest sensible retention and the fewest people who need access. Prove the accuracy in the real environment, publish a clear explanation for affected people and keep human judgement where a missed or false event could matter. That is a better foundation than buying the most advanced sensing feature because the word “Wi-Fi” makes it sound free.

Further reading

Wi-Fi sensing is developing quickly. Standards, product capabilities, spectrum use, supplier terms and UK privacy guidance may change. Check the current documentation before making a deployment decision.

What Should Your IT Provider Be Monitoring? →

Why CAPTCHAs Are Being Replaced — And Why Business Owners Should Care →