The Last Job the Elevator Operator Ever Had
There was a time when every elevator needed an operator.
Early elevator systems weren't designed for independence. They required constant human control to function safely and effectively. An operator managed the speed of the cab, aligned it precisely with each floor, and handled the opening and closing of the doors. Every movement depended on that person's attention and skill. Without them, the system didn't work.
That model held until the system itself evolved.
Push-button controls began to appear in the early 1900s, allowing passengers to select their destination directly. Over time, improvements in control systems, safety mechanisms, and automation made it possible for elevators to manage these tasks on their own. As the technology matured, the need for manual execution disappeared. By the mid-20th century, automated elevators were the standard. Within a few decades, operators were no longer part of the system.
No one looks back on that transition as a loss.
The system became more consistent, more efficient, and always available. It no longer depended on presence or performance. Execution moved into the system, and everything around it improved.
That same pattern is now emerging in physical security.
You Can See Everything. You Can't Act on All of It.
Most organizations already have visibility across their environments.
Cameras are deployed across perimeters, parking areas, entry points, and interior spaces. Over the past decade, significant investment has gone into expanding that visibility. As a result, most enterprises can see far more than they ever could before. The infrastructure is there, and in many cases, it's already delivering high-quality detection and analytics.
What hasn't kept pace is execution.
Monitoring those systems in real time requires people, and there simply aren't enough of them to match the scale of modern deployments. Operators are often responsible for dozens or even hundreds of feeds simultaneously. In many environments, a large percentage of cameras aren't actively monitored at all. They serve a forensic purpose, allowing teams to review what happened after an incident has already occurred.
As environments scale, this gap becomes unavoidable. Visibility expands. Response remains constrained by time, attention, and staffing.
That's the inflection point.
The question is no longer whether organizations can detect activity. It's whether they can act on it consistently, in real time, across every site and every system.
Automating the Patrol
This challenge becomes more pronounced in large or distributed environments.
Perimeters extend across wide areas. Activity is spread across multiple zones. Guard tours are designed to provide coverage, but they require time to complete and depend on schedules, routing, and availability. Even well-executed patrols introduce gaps because they're sequential by nature. A guard can only be in one place at a time.
Automation changes how that coverage is delivered.
Autonomous patrol units move continuously across large environments without interruption. Edge-based systems detect and respond to activity the moment it occurs, directly at the source. Coverage becomes persistent rather than periodic, and awareness is no longer tied to a specific moment.
This doesn't remove personnel from the operation. It gives them better starting points.
Instead of covering ground and searching for activity, teams respond to verified events with clear context around what's happening and where attention is needed. The result is a more effective use of resources and a sharper level of operational awareness.
The Alert Came In. Now What?
Inside the security operations center, a similar shift is taking place.
The traditional model is built around a linear workflow. An alert is generated, an operator reviews it, a decision is made, and action follows. This approach works within a controlled volume, but as alert volume increases, consistency becomes harder to maintain. False alarms consume time. Multiple events compete for attention. Prioritization becomes constant.
Even experienced teams run into limits under these conditions, not because of capability, but because the model depends on sequential human execution.
What's changing now is the role of the system.
Modern platforms go beyond detection and alerting. They verify events, initiate deterrence, notify stakeholders, and document the incident as it unfolds. These actions don't need to wait, and they don't need to happen one at a time. They execute simultaneously, across multiple channels.
This is where Agentic AI operates.
AI agents don't just surface information. They work the incident. They investigate what's happening, determine whether it requires action, and execute response protocols in real time. Voice deterrence, notifications, escalation, and reporting are coordinated as a single process.
Execution moves into the system.
From Operator to Orchestration
This transition reflects how complex systems evolve.
In early elevator systems, the operator handled every step. Today, that responsibility sits within the system itself. People still interact with it, but at a higher level, focused on outcomes rather than manual control.
Security is moving along that same path.
Detection is already automated. Verification continues to improve. Response is now moving into the system.
This shift allows teams to focus on higher-value work. Oversight, decision-making, and intervention remain critical, but they're supported by systems already executing the workflow in real time.
Incident Response at Scale
At scale, response isn't a single action. It's a coordinated set of actions across systems, people, and environments.
In manual workflows, these steps happen one at a time. That introduces delay and variability.
An orchestrated model brings those steps together.
Agentic AI connects detection, verification, deterrence, escalation, and documentation into a unified process. Actions that were previously sequential now occur simultaneously. Voice deterrence can begin while stakeholders are notified. Incident data is captured and structured while the event is still unfolding.
This level of coordination delivers consistency across every site and every scenario.
This Isn't Coming
This model is already operating in real environments.
With SARA, RAD's Agentic AI, response begins the moment an event is detected. The system evaluates the situation, confirms whether it requires action, and initiates response immediately.
Voice deterrence is delivered in real time, tailored to the situation. Stakeholders are notified with relevant context. Escalation is initiated without delay when required. At the same time, a complete incident record is built, including video, audio, and timestamps.
This isn't a workflow improvement on paper.
It's a measurable change in how incidents are handled.
Response begins earlier. Actions are applied consistently. Outcomes improve because execution no longer depends on availability.
Security teams focus on verified incidents, with full visibility into what's already occurred and what actions are already in motion.
What Comes Next
Every system reaches a point where manual execution limits performance.
Automation expands capability. It increases coverage, improves consistency, and accelerates response without requiring proportional increases in resources.
Physical security has reached that point.
The industry has invested in visibility. Detection capabilities have advanced. The next stage is execution.
Elevator operators were essential when systems required manual control. As the system evolved, execution moved into the machine, and the entire experience improved.
Security is moving through that same transition.
Execution is shifting into systems that can investigate, triage, and act in real time, across environments, without waiting. This allows teams to operate with greater awareness, apply expertise where it matters most, and deliver a level of response that manual systems alone can't sustain.
That's where the industry is headed.
And in many environments, it's already here.
David Marsh Vice President of Marketing Robotic Assistance Devices linkedin.com/in/davidmarsh
To learn how SARA Agentic AI helps security teams strengthen security operations from detection to resolution, visit radsecurity.com/sara.
