Steve Lindsey, Chief Technology Officer at LiveView Technologies (LVT) examines the pivotal impact of Agentic AI on the security industry.
The physical security industry is rooted in reactionary measures that require humans to respond.
Teams monitor alert dashboards and rush to address incidents if someone is available.
Otherwise, they analyze evidence afterward and attempt to pursue bad actors.
Because of these highly manual processes – reviewing alerts from disconnected systems, prioritizing incidents in real time, sifting through hours of footage – security coverage is often lacking.
Simply put, physical security is constrained by limited human resources.
Human guards can cover only so much ground and every security technology investment requires an ever-increasing upkeep budget.
Ensuring 24/7 security can feel impossible, forcing teams to make tradeoffs considering their priorities and resources.
But this no longer has to be the reality thanks to AI.
AI has evolved to a point where it enables proactive security management.
Instead of passively observing surroundings and collecting data for future analysis, AI can work in real time to combine insights across any security investment, analyze threats and execute the appropriate response at that moment.
Agentic AI, specifically, is the key enabler in this shift.
Given that the technology is nascent, many organizations do not yet realize its power and potential – or what it takes to successfully integrate Agentic AI into a security strategy.
As a technology leader with decades of experience building AI-powered video intelligence systems, I’ve heard from countless security leaders who are confused about the state of AI in security and where Agentic AI fits.
To help your organization set the foundation for future Agentic AI implementations, let’s explore what Agentic AI is, how the technology will reinvent the security industry and what organizations need to implement it.
Agentic AI, as its name suggests, can act in intelligent ways similar to human agents and orchestrate a complete process that is both dynamic and designed to achieve an organizational goal and outcome.
In the context of physical security, Agentic AI leverages insights from every security device to gain intelligence and situational awareness and, in turn, execute the most appropriate response to threats or situations that it detects.
Agentic AI builds on the existing AI foundations that have evolved greatly in recent years:
These advancements in AI underpin Agentic AI, so it can perform in ways indistinguishable from actual humans.
Agentic AI is unique in its ability to make intelligent decisions and take action dynamically, instead of more traditional AI like computer vision models that are designed to simply yield outputs (alerts) based on a specific object it is trained to detect (like a person, vehicle or animal).
To activate Agentic AI for security effectively, it needs:
Humans will always govern agentic systems and Agentic AI greatly extends what humans are capable of doing.
Human security teams have full control to monitor and adjust the system, ensuring it responds to situations appropriately and does not act in ways that conflict with the organization’s values.
Agentic AI is highly effective for filtering through the noise that security teams face.
Every camera, sensor and other device creates data and alerts that quickly become noise for humans, causing alert fatigue.
Agentic AI understands how to assess alerts, implement real-time security measures where appropriate and identify which situations call for human attention or intervention.
Agentic AI removes the need for customized software development or integrations to build connected security ecosystems.
Agentic systems can access the data and functions they need to perform various tasks, such as:
Detecting abnormal behavior in context: Identify possible bad actors in an environment, leveraging cameras and detection technologies to know if people are trespassing, acting in ways that may indicate criminal activity or are in life-threatening situations.
A very important difference between traditional computer vision AI and Agentic AI in detection is that Agentic AI can consider behavior and context in ways that traditional computer vision cannot.
This further reduces noise and highlights real problems or concerns that should be focused on.
Effectively deter threats or modify behavior in real-time: Deploy human-like deterrence capabilities to ward off bad actors before they commit a crime and remind employees of behavior that could be life-threatening to themselves or others.
Agentic AI can utilize loudspeakers, spotlights or alarms in effective and believable ways to deter would-be criminals (and possibly stop them from ever coming back) or to modify behavior that can actually prevent problems from happening.
Gather and package evidence in real-time: Effective deterrence is 100% reliant on holding people accountable. Accountability is reliant on effective prosecution.
For years, prosecution has been expensive in both time and money, which has made prosecution more and more rare and criminals increasingly brazen.
If criminals ignore the human-like deterrence measures, the Agentic AI can automatically gather and package evidence of an incident as it occurs, tagging the video files and marking information that will accelerate an investigation.
Evidence can be easily found using natural language search terms and descriptions on any video from any camera(s), from any time.
Faster evidence gathering lowers time and cost to prosecute, resulting in more cases being tried and ultimately more accountability and more effectiveness of deterrence.
Agentic AI is a force multiplier that maximizes the value of any security investment while empowering human professionals with the information they need and helping them focus on the most important issues that require their attention.
With Agentic AI, every camera, sensor and security tool can work in perfect harmony, significantly accelerating response times to any incident – and maximizing the time and cost efficiency of prosecution if needed.
The technology allows for agility and future-proofing like we’ve never seen before. As criminals put countermeasures in place, the Agentic AI reacts to those countermeasures in real time so teams are always one step ahead.
Agentic AI will forever change the security industry and we’ve just begun to see its potential.
These systems will rapidly improve as they continue to learn from each environment and activate increasingly powerful cloud and edge devices designed for any security need.
Humans can partner with AI systems to address an increasingly wide range of threats and ensure every incident receives the appropriate attention, including an efficient prosecution process when crimes occur.
Through the power of AI and human professionals, businesses across industries can run highly efficient security processes, maximize safety for employees and consumers, and send a clear message to would-be criminals that they will be stopped.
This article was originally published in the June edition of Security Journal Americas. To read your FREE digital edition, click here.