AI computer vision for lift and elevator safety monitoring

The problem: nobody sees the lift until it fails

Lift incidents are usually discovered late. A passenger is trapped between floors and relies on an intercom that may not be answered. A door stalls repeatedly for weeks before anyone raises a complaint. A goods lift is overloaded every evening because it is quicker than making two trips. In each case the information existed on camera, but no one was watching, and the footage was only reviewed after something went badly wrong.

Manual controls do not close this gap. A maintenance visit sees the lift on one morning a month. A security guard watching a wall of screens cannot reliably notice a door that closed slightly too slowly. And in most buildings, the person accountable for lift safety — the facility manager or society committee — has no continuous view at all.

How our solution works

The system reads video from cameras already installed in the cabin, lobby or machine room. A vision model tracks people and doors frame by frame and compares what it sees against rules that are set up for your building: how long a door may remain open, how long a cabin may stay still with a person inside, how many people the cabin should hold, whether anyone is at the door line as it closes.

When a rule is breached, the system classifies the event and routes it accordingly. Entrapment, a passenger who has fallen, or a person stuck in a closing door are treated as live alerts to the security desk and facility manager, with a short clip attached so the responder can see the situation before they move. Slower-burning signals — doors that stall, unusual vibration patterns, repeated overloading — are collected into trend reports that feed maintenance planning instead of interrupting anyone.

Because these patterns are rules over time rather than one-off snapshots, the system also becomes an evidence trail: which lift, which hour, how often. That is typically what changes a maintenance contract conversation from opinion into fact.

What is delivered

• ✓Real-time detection of entrapment, falls, door obstruction and overcrowding
• ✓Live alerts to the security desk and facility manager with a short event clip
• ✓Anomaly trends such as repeated door stalls or recurring overload by lift and shift
• ✓A dashboard by building, lift and time period for maintenance planning
• ✓Event log with retention limits and role-based access for compliance review
• ✓Configuration workshop to define what counts as an incident in your building

Who it is for

Facility management companies, commercial building and mall operators, hospitals, hotels, large residential societies, and lift maintenance firms who want to offer monitoring alongside servicing. It is most valuable where lifts run heavy traffic, where vulnerable passengers use them, or where the operator manages many buildings and cannot be present at any of them.

Buildings with a single low-traffic lift and an attentive on-site team may find manual checks sufficient. Where there are several lifts across several sites, the economics change quickly — one avoided entrapment or one correctly predicted door failure is usually worth more than the monitoring itself.

Frequently asked questions